2025
|
Njoku, Uchechukwu U. F.; Abelló, Alberto; Bilalli, Besim; Bontempi, Gianluca On many-objective feature selection and the need for interpretability Journal Article In: Expert systems with applications, vol. 267, 2025, (DOI: 10.1016/j.eswa.2024.126191). @article{info:hdl:2013/388988b,
title = {On many-objective feature selection and the need for interpretability},
author = {Uchechukwu U. F. Njoku and Alberto Abelló and Besim Bilalli and Gianluca Bontempi},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/388988/1/doi_372632.pdf},
year = {2025},
date = {2025-01-01},
journal = {Expert systems with applications},
volume = {267},
abstract = {Big data comes with the challenge of containing irrelevant and redundant information (i.e., features). Given that a single objective cannot fully capture a feature's relevance, a Many-Objective Feature Selection (MOFS) approach able to accommodate various relevant perspectives is preferred for identifying the most appropriate features in a given context. However, MOFS produces a large set of solutions whose interpretability has been largely overlooked. First, we demonstrate the relevance of MOFS and establish its necessity by considering up to six objectives using a genetic algorithm and Naive Bayes on ten datasets for classification tasks. Then, we propose a novel methodology to improve the interpretability of MOFS results in order to support the data scientist in selecting the subset of features pertinent to their use case. Our methodology is instantiated as an intuitive and interactive dashboard that provides insights into the results beyond the pure numerical representation of the objectives being considered and evaluated with 50 participants. The outcome shows that it addresses the need for a methodological approach and comprehensive visualization to achieve interoperability.},
note = {DOI: 10.1016/j.eswa.2024.126191},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Big data comes with the challenge of containing irrelevant and redundant information (i.e., features). Given that a single objective cannot fully capture a feature's relevance, a Many-Objective Feature Selection (MOFS) approach able to accommodate various relevant perspectives is preferred for identifying the most appropriate features in a given context. However, MOFS produces a large set of solutions whose interpretability has been largely overlooked. First, we demonstrate the relevance of MOFS and establish its necessity by considering up to six objectives using a genetic algorithm and Naive Bayes on ten datasets for classification tasks. Then, we propose a novel methodology to improve the interpretability of MOFS results in order to support the data scientist in selecting the subset of features pertinent to their use case. Our methodology is instantiated as an intuitive and interactive dashboard that provides insights into the results beyond the pure numerical representation of the objectives being considered and evaluated with 50 participants. The outcome shows that it addresses the need for a methodological approach and comprehensive visualization to achieve interoperability. |
Njoku, Uchechukwu U. F.; Abelló, Alberto; Bilalli, Besim; Bontempi, Gianluca On many-objective feature selection and the need for interpretability Journal Article In: Expert systems with applications, vol. 267, 2025, (DOI: 10.1016/j.eswa.2024.126191). @article{info:hdl:2013/388988,
title = {On many-objective feature selection and the need for interpretability},
author = {Uchechukwu U. F. Njoku and Alberto Abelló and Besim Bilalli and Gianluca Bontempi},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/388988/1/doi_372632.pdf},
year = {2025},
date = {2025-01-01},
journal = {Expert systems with applications},
volume = {267},
abstract = {Big data comes with the challenge of containing irrelevant and redundant information (i.e., features). Given that a single objective cannot fully capture a feature's relevance, a Many-Objective Feature Selection (MOFS) approach able to accommodate various relevant perspectives is preferred for identifying the most appropriate features in a given context. However, MOFS produces a large set of solutions whose interpretability has been largely overlooked. First, we demonstrate the relevance of MOFS and establish its necessity by considering up to six objectives using a genetic algorithm and Naive Bayes on ten datasets for classification tasks. Then, we propose a novel methodology to improve the interpretability of MOFS results in order to support the data scientist in selecting the subset of features pertinent to their use case. Our methodology is instantiated as an intuitive and interactive dashboard that provides insights into the results beyond the pure numerical representation of the objectives being considered and evaluated with 50 participants. The outcome shows that it addresses the need for a methodological approach and comprehensive visualization to achieve interoperability.},
note = {DOI: 10.1016/j.eswa.2024.126191},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Big data comes with the challenge of containing irrelevant and redundant information (i.e., features). Given that a single objective cannot fully capture a feature's relevance, a Many-Objective Feature Selection (MOFS) approach able to accommodate various relevant perspectives is preferred for identifying the most appropriate features in a given context. However, MOFS produces a large set of solutions whose interpretability has been largely overlooked. First, we demonstrate the relevance of MOFS and establish its necessity by considering up to six objectives using a genetic algorithm and Naive Bayes on ten datasets for classification tasks. Then, we propose a novel methodology to improve the interpretability of MOFS results in order to support the data scientist in selecting the subset of features pertinent to their use case. Our methodology is instantiated as an intuitive and interactive dashboard that provides insights into the results beyond the pure numerical representation of the objectives being considered and evaluated with 50 participants. The outcome shows that it addresses the need for a methodological approach and comprehensive visualization to achieve interoperability. |
2024
|
Terrucha, Ines; Domingos, Elias Fernandez; Suchon, Remi; Santos, Francisco C; Simoens, Pieter; Lenaerts, Tom Humans program artificial delegates to accurately solve collective-risk dilemmas, but lack precision Miscellaneous 2024, (Conference: Machine+behavior Conference(Berlin, Allemagne)). @misc{info:hdl:2013/385912b,
title = {Humans program artificial delegates to accurately solve collective-risk dilemmas, but lack precision},
author = {Ines Terrucha and Elias Fernandez Domingos and Remi Suchon and Francisco C Santos and Pieter Simoens and Tom Lenaerts},
url = {http://hdl.handle.net/2013/ULB-DIPOT:oai:dipot.ulb.ac.be:2013/385912},
year = {2024},
date = {2024-01-01},
note = {Conference: Machine+behavior Conference(Berlin, Allemagne)},
keywords = {},
pubstate = {published},
tppubtype = {misc}
}
|
Kirchsteiger, Georg; Lenaerts, Tom; Suchon, Remi Growing cooperation Miscellaneous 2024, (Conference: Conference of the French Experimental Economics Association(14: grenoble, France)). @misc{info:hdl:2013/385911b,
title = {Growing cooperation},
author = {Georg Kirchsteiger and Tom Lenaerts and Remi Suchon},
url = {http://hdl.handle.net/2013/ULB-DIPOT:oai:dipot.ulb.ac.be:2013/385911},
year = {2024},
date = {2024-01-01},
note = {Conference: Conference of the French Experimental Economics Association(14: grenoble, France)},
keywords = {},
pubstate = {published},
tppubtype = {misc}
}
|
Bosch, Inas; Gravel, Barbara; Lenaerts, Tom Knowledge graph embeddings for the prediction of pathogenic gene pairs Miscellaneous 2024, (Conference: European Conference on Computational Biology.(23: 16/09-20/09/2024: Turku, Finland)). @misc{info:hdl:2013/385910b,
title = {Knowledge graph embeddings for the prediction of pathogenic gene pairs},
author = {Inas Bosch and Barbara Gravel and Tom Lenaerts},
url = {http://hdl.handle.net/2013/ULB-DIPOT:oai:dipot.ulb.ac.be:2013/385910},
year = {2024},
date = {2024-01-01},
note = {Conference: European Conference on Computational Biology.(23: 16/09-20/09/2024: Turku, Finland)},
keywords = {},
pubstate = {published},
tppubtype = {misc}
}
|
Gravel, Barbara; Renaux, Alexandre; Papadimitriou, Sofia; Smits, Guillaume; Nowé, Ann; Lenaerts, Tom Prioritization of variant combinations in whole exomes Miscellaneous 2024, (Conference: European Conference on Computational Biology.(23: 16/09-20/09/2024: Turku, Finland)). @misc{info:hdl:2013/385909b,
title = {Prioritization of variant combinations in whole exomes},
author = {Barbara Gravel and Alexandre Renaux and Sofia Papadimitriou and Guillaume Smits and Ann Nowé and Tom Lenaerts},
url = {http://hdl.handle.net/2013/ULB-DIPOT:oai:dipot.ulb.ac.be:2013/385909},
year = {2024},
date = {2024-01-01},
note = {Conference: European Conference on Computational Biology.(23: 16/09-20/09/2024: Turku, Finland)},
keywords = {},
pubstate = {published},
tppubtype = {misc}
}
|
Abels, Axel; Lenaerts, Tom; Trianni, Vito; Nowé, Ann Dealing with Expert Bias in Collective Decision-making Miscellaneous 2024, (Conference: European Conference on Artificial Intelligence(27: 19/10-24/10/2024: Santiago de Compostella)). @misc{info:hdl:2013/385908b,
title = {Dealing with Expert Bias in Collective Decision-making},
author = {Axel Abels and Tom Lenaerts and Vito Trianni and Ann Nowé},
url = {http://hdl.handle.net/2013/ULB-DIPOT:oai:dipot.ulb.ac.be:2013/385908},
year = {2024},
date = {2024-01-01},
note = {Conference: European Conference on Artificial Intelligence(27: 19/10-24/10/2024: Santiago de Compostella)},
keywords = {},
pubstate = {published},
tppubtype = {misc}
}
|
Leung, Chin Wing; Lenaerts, Tom; Turrini, Paolo To Promote Full Cooperation in Social Dilemmas, Agents Need to Unlearn Loyalty Proceedings Article In: Larson, Kate (Ed.): Proceedings of the Thirty-Third International Joint Conference on Artificial Intelligence, pp. 111-119, International Joint Conferences on Artificial Intelligence (IJCAI) Organization, 2024, (Conference: International Joint Conference on Artificial Intelligence(33: 3/8-9/8/2024: Jeju. Korea)). @inproceedings{info:hdl:2013/385907b,
title = {To Promote Full Cooperation in Social Dilemmas, Agents Need to Unlearn Loyalty},
author = {Chin Wing Leung and Tom Lenaerts and Paolo Turrini},
editor = {Kate Larson},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/385907/3/0013.pdf},
year = {2024},
date = {2024-01-01},
booktitle = {Proceedings of the Thirty-Third International Joint Conference on Artificial Intelligence},
pages = {111-119},
publisher = {International Joint Conferences on Artificial Intelligence (IJCAI) Organization},
abstract = {If given the choice, what strategy should agents use to switch partners in strategic social interactions? While many analyses have been performed on specific switching heuristics, showing how and when these lead to more cooperation, no insights have been provided into which rule will actually be learnt by agents when given the freedom to do so. Starting from a baseline model that has demonstrated the potential of rewiring for cooperation, we provide answers to this question over the full spectrum of social dilemmas. Multi-agent Q-learning with Boltzmann exploration is used to learn when to sever or maintain an association. In both the Prisoner's Dilemma and the Stag Hunt games we observe that the Out-for-Tat rewiring rule, breaking ties with other agents choosing socially undesirable actions, becomes dominant, confirming at the same time that cooperation flourishes when rewiring is fast enough relative to imitation. Nonetheless, in the transitory region before full cooperation, a Stay strategy, keeping a connection at all costs, remains present, which shows that loyalty needs to be overcome for full cooperation to emerge. In conclusion, individuals learn cooperation-promoting rewiring rules but need to overcome a kind of loyalty to achieve full cooperation in the full spectrum of social dilemmas.},
note = {Conference: International Joint Conference on Artificial Intelligence(33: 3/8-9/8/2024: Jeju. Korea)},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
If given the choice, what strategy should agents use to switch partners in strategic social interactions? While many analyses have been performed on specific switching heuristics, showing how and when these lead to more cooperation, no insights have been provided into which rule will actually be learnt by agents when given the freedom to do so. Starting from a baseline model that has demonstrated the potential of rewiring for cooperation, we provide answers to this question over the full spectrum of social dilemmas. Multi-agent Q-learning with Boltzmann exploration is used to learn when to sever or maintain an association. In both the Prisoner's Dilemma and the Stag Hunt games we observe that the Out-for-Tat rewiring rule, breaking ties with other agents choosing socially undesirable actions, becomes dominant, confirming at the same time that cooperation flourishes when rewiring is fast enough relative to imitation. Nonetheless, in the transitory region before full cooperation, a Stay strategy, keeping a connection at all costs, remains present, which shows that loyalty needs to be overcome for full cooperation to emerge. In conclusion, individuals learn cooperation-promoting rewiring rules but need to overcome a kind of loyalty to achieve full cooperation in the full spectrum of social dilemmas. |
Molinghen, Yannick; Avalos, Raphaël; Achter, Mark Van; Nowé, Ann; Lenaerts, Tom Laser Learning Environment: A new environment for coordination-critical multi-agent tasks Proceedings Article In: Oliehoek, Frans F. A.; Manon, Kok; Verwer, Sicco (Ed.): Artificial Intelligence and Machine Learning: Revised Selected Papers, Springer Science and Business Media Deutschland GmbH, 2024, (Conference: Benelux Conference Ai conference, BNAIC(35: 8-10/11/2023: TU Delft)). @inproceedings{info:hdl:2013/370546b,
title = {Laser Learning Environment: A new environment for coordination-critical multi-agent tasks},
author = {Yannick Molinghen and Raphaël Avalos and Mark Van Achter and Ann Nowé and Tom Lenaerts},
editor = {Frans F. A. Oliehoek and Kok Manon and Sicco Verwer},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/370546/4/2404.03596v1.pdf},
year = {2024},
date = {2024-01-01},
booktitle = {Artificial Intelligence and Machine Learning: Revised Selected Papers},
publisher = {Springer Science and Business Media Deutschland GmbH},
series = {Communications in Computer and Information Science},
abstract = {We introduce the Laser Learning Environment (LLE), a collaborative multi-agent reinforcement learning environment where coordination is key. In LLE, agents depend on each other to make progress (interdependence), must jointly take specific sequences of actions to succeed (perfect coordination), and accomplishing those joint actions does not yield any intermediate reward (zero-incentive dynamics). The challenge of such problems lies in the difficulty of escaping state space bottlenecks caused by interdependence steps since escaping those bottlenecks is not rewarded. We test multiple state-of-the-art value-based MARL algorithms against LLE and show that they consistently fail at the collaborative task because of their inability to escape state space bottlenecks, even though they successfully achieve perfect coordination. We show that Q-learning extensions such as prioritised experience replay and n-steps return hinder exploration in environments with zero-incentive dynamics, and find that intrinsic curiosity with random network distillation is not sufficient to escape those bottlenecks. We demonstrate the need for novel methods to solve this problem and the relevance of LLE as cooperative MARL benchmark.},
note = {Conference: Benelux Conference Ai conference, BNAIC(35: 8-10/11/2023: TU Delft)},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
We introduce the Laser Learning Environment (LLE), a collaborative multi-agent reinforcement learning environment where coordination is key. In LLE, agents depend on each other to make progress (interdependence), must jointly take specific sequences of actions to succeed (perfect coordination), and accomplishing those joint actions does not yield any intermediate reward (zero-incentive dynamics). The challenge of such problems lies in the difficulty of escaping state space bottlenecks caused by interdependence steps since escaping those bottlenecks is not rewarded. We test multiple state-of-the-art value-based MARL algorithms against LLE and show that they consistently fail at the collaborative task because of their inability to escape state space bottlenecks, even though they successfully achieve perfect coordination. We show that Q-learning extensions such as prioritised experience replay and n-steps return hinder exploration in environments with zero-incentive dynamics, and find that intrinsic curiosity with random network distillation is not sufficient to escape those bottlenecks. We demonstrate the need for novel methods to solve this problem and the relevance of LLE as cooperative MARL benchmark. |
Attafi, Omar Abdelghani; Clementel, Damiano; Kyritsis, Konstantinos; Capriotti, Emidio; Farrell, Gavin; Fragkouli, Styliani-Christina; Castro, Leyla Jael; Hatos, András; Lenaerts, Tom; Mazurenko, Stanislav; Mozaffari, Soroush; Pradelli, Franco; Ruch, Patrick; Savojardo, Castrense; Turina, Maria Paola; Zambelli, Federico; Piovesan, Damiano; Monzon, Alexander Miguel; Psomopoulos, Fotis F. E.; Tosatto, Silvio S. C. E. DOME Registry: implementing community-wide recommendations for reporting supervised machine learning in biology Journal Article In: GigaScience, vol. 13, pp. 8, 2024, (DOI: 10.1093/gigascience/giae094). @article{info:hdl:2013/385906b,
title = {DOME Registry: implementing community-wide recommendations for reporting supervised machine learning in biology},
author = {Omar Abdelghani Attafi and Damiano Clementel and Konstantinos Kyritsis and Emidio Capriotti and Gavin Farrell and Styliani-Christina Fragkouli and Leyla Jael Castro and András Hatos and Tom Lenaerts and Stanislav Mazurenko and Soroush Mozaffari and Franco Pradelli and Patrick Ruch and Castrense Savojardo and Maria Paola Turina and Federico Zambelli and Damiano Piovesan and Alexander Miguel Monzon and Fotis F. E. Psomopoulos and Silvio S. C. E. Tosatto},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/385906/3/giae094-2.pdf},
year = {2024},
date = {2024-01-01},
journal = {GigaScience},
volume = {13},
pages = {8},
abstract = {Abstract Supervised machine learning (ML) is used extensively in biology and deserves closer scrutiny. The Data Optimization Model Evaluation (DOME) recommendations aim to enhance the validation and reproducibility of ML research by establishing standards for key aspects such as data handling and processing, optimization, evaluation, and model interpretability. The recommendations help to ensure that key details are reported transparently by providing a structured set of questions. Here, we introduce the DOME registry (URL: registry.dome-ml.org), a database that allows scientists to manage and access comprehensive DOME-related information on published ML studies. The registry uses external resources like ORCID, APICURON, and the Data Stewardship Wizard to streamline the annotation process and ensure comprehensive documentation. By assigning unique identifiers and DOME scores to publications, the registry fosters a standardized evaluation of ML methods. Future plans include continuing to grow the registry through community curation, improving the DOME score definition and encouraging publishers to adopt DOME standards, and promoting transparency and reproducibility of ML in the life sciences.},
note = {DOI: 10.1093/gigascience/giae094},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Abstract Supervised machine learning (ML) is used extensively in biology and deserves closer scrutiny. The Data Optimization Model Evaluation (DOME) recommendations aim to enhance the validation and reproducibility of ML research by establishing standards for key aspects such as data handling and processing, optimization, evaluation, and model interpretability. The recommendations help to ensure that key details are reported transparently by providing a structured set of questions. Here, we introduce the DOME registry (URL: registry.dome-ml.org), a database that allows scientists to manage and access comprehensive DOME-related information on published ML studies. The registry uses external resources like ORCID, APICURON, and the Data Stewardship Wizard to streamline the annotation process and ensure comprehensive documentation. By assigning unique identifiers and DOME scores to publications, the registry fosters a standardized evaluation of ML methods. Future plans include continuing to grow the registry through community curation, improving the DOME score definition and encouraging publishers to adopt DOME standards, and promoting transparency and reproducibility of ML in the life sciences. |
Juchnewitsch, Anna Grete; Pomm, Kristjan; Dutta, Avirup; Tamp, Erik; Valkna, Anu; Lillepea, Kristiina; Mahyari, Eisa; Tjagur, Stanislav; Belova, Galina; Kübarsepp, Viljo; Castillo-Madeen, Helen; Riera-Escamilla, Antoni; Põlluaas, Lisanna; Nagirnaja, Liina; Poolamets, Olev; Vihljajev, Vladimir; Sütt, Mailis; Versbraegen, Nassim; Papadimitriou, Sofia; McLachlan, Robert Ian; Jarvi, Keith Allen; Schlegel, Peter P. N.; Tennisberg, Sven; Korrovits, Paul; Vigh-Conrad, Katinka; O’Bryan, Moira M. K.; Aston, Kenneth Ivan; Lenaerts, Tom; Conrad, Donald D. F.; Kasak, Laura; Punab, Margus; Laan, Maris Undiagnosed RASopathies in infertile men Journal Article In: Frontiers in endocrinology, vol. 15, 2024, (DOI: 10.3389/fendo.2024.1312357). @article{info:hdl:2013/374860b,
title = {Undiagnosed RASopathies in infertile men},
author = {Anna Grete Juchnewitsch and Kristjan Pomm and Avirup Dutta and Erik Tamp and Anu Valkna and Kristiina Lillepea and Eisa Mahyari and Stanislav Tjagur and Galina Belova and Viljo Kübarsepp and Helen Castillo-Madeen and Antoni Riera-Escamilla and Lisanna Põlluaas and Liina Nagirnaja and Olev Poolamets and Vladimir Vihljajev and Mailis Sütt and Nassim Versbraegen and Sofia Papadimitriou and Robert Ian McLachlan and Keith Allen Jarvi and Peter P. N. Schlegel and Sven Tennisberg and Paul Korrovits and Katinka Vigh-Conrad and Moira M. K. O’Bryan and Kenneth Ivan Aston and Tom Lenaerts and Donald D. F. Conrad and Laura Kasak and Margus Punab and Maris Laan},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/374860/1/doi_358504.pdf},
year = {2024},
date = {2024-01-01},
journal = {Frontiers in endocrinology},
volume = {15},
abstract = {RASopathies are syndromes caused by congenital defects in the Ras/mitogen-activated protein kinase (MAPK) pathway genes, with a population prevalence of 1 in 1,000. Patients are typically identified in childhood based on diverse characteristic features, including cryptorchidism (CR) in >50% of affected men. As CR predisposes to spermatogenic failure (SPGF; total sperm count per ejaculate 0–39 million), we hypothesized that men seeking infertility management include cases with undiagnosed RASopathies. Likely pathogenic or pathogenic (LP/P) variants in 22 RASopathy-linked genes were screened in 521 idiopathic SPGF patients (including 155 CR cases) and 323 normozoospermic controls using exome sequencing. All 844 men were recruited to the ESTonian ANDrology (ESTAND) cohort and underwent identical andrological phenotyping. RASopathy-specific variant interpretation guidelines were used for pathogenicity assessment. LP/P variants were identified in PTPN11 (two), SOS1 (three), SOS2 (one), LZTR1 (one), SPRED1 (one), NF1 (one), and MAP2K1 (one). The findings affected six of 155 cases with CR and SPGF, three of 366 men with SPGF only, and one (of 323) normozoospermic subfertile man. The subgroup “CR and SPGF” had over 13-fold enrichment of findings compared to controls (3.9% vs. 0.3%; Fisher’s exact test},
note = {DOI: 10.3389/fendo.2024.1312357},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
RASopathies are syndromes caused by congenital defects in the Ras/mitogen-activated protein kinase (MAPK) pathway genes, with a population prevalence of 1 in 1,000. Patients are typically identified in childhood based on diverse characteristic features, including cryptorchidism (CR) in >50% of affected men. As CR predisposes to spermatogenic failure (SPGF; total sperm count per ejaculate 0–39 million), we hypothesized that men seeking infertility management include cases with undiagnosed RASopathies. Likely pathogenic or pathogenic (LP/P) variants in 22 RASopathy-linked genes were screened in 521 idiopathic SPGF patients (including 155 CR cases) and 323 normozoospermic controls using exome sequencing. All 844 men were recruited to the ESTonian ANDrology (ESTAND) cohort and underwent identical andrological phenotyping. RASopathy-specific variant interpretation guidelines were used for pathogenicity assessment. LP/P variants were identified in PTPN11 (two), SOS1 (three), SOS2 (one), LZTR1 (one), SPRED1 (one), NF1 (one), and MAP2K1 (one). The findings affected six of 155 cases with CR and SPGF, three of 366 men with SPGF only, and one (of 323) normozoospermic subfertile man. The subgroup “CR and SPGF” had over 13-fold enrichment of findings compared to controls (3.9% vs. 0.3%; Fisher’s exact test |
Stefanija, Ana Pop; Buelens, Bart; Goesaert, Elfi; Lenaerts, Tom; Pierson, Jean René; den Bussche, Jan Van Toward a Solid Acceptance of the Decentralized Web of Personal Data: Societal and Technological Convergence Journal Article In: Communications of the ACM, vol. 67, no. 1, pp. 43-46, 2024, (DOI: 10.1145/3624555). @article{info:hdl:2013/367025b,
title = {Toward a Solid Acceptance of the Decentralized Web of Personal Data: Societal and Technological Convergence},
author = {Ana Pop Stefanija and Bart Buelens and Elfi Goesaert and Tom Lenaerts and Jean René Pierson and Jan Van den Bussche},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/367025/4/3624555.pdf},
year = {2024},
date = {2024-01-01},
journal = {Communications of the ACM},
volume = {67},
number = {1},
pages = {43-46},
abstract = {Giving individuals more control of their personal data.},
note = {DOI: 10.1145/3624555},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Giving individuals more control of their personal data. |
Lenaerts, Tom; Saponara, Marco; Pacheco, Jorge J. M.; Santos, Francisco C. Evolution of a theory of mind Journal Article In: iScience, vol. 27, no. 2, 2024, (DOI: 10.1016/j.isci.2024.108862). @article{info:hdl:2013/372022b,
title = {Evolution of a theory of mind},
author = {Tom Lenaerts and Marco Saponara and Jorge J. M. Pacheco and Francisco C. Santos},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/372022/1/doi_355666.pdf},
year = {2024},
date = {2024-01-01},
journal = {iScience},
volume = {27},
number = {2},
abstract = {Even though the Theory of Mind in upper primates has been under investigation for decades, how it may evolve remains an open problem. We propose here an evolutionary game theoretical model where a finite population of individuals may use reasoning strategies to infer a response to the anticipated behavior of others within the context of a sequential dilemma, i.e., the Centipede Game. We show that strategies with bounded reasoning evolve and flourish under natural selection, provided they are allowed to make reasoning mistakes and a temptation for higher future gains is in place. We further show that non-deterministic reasoning co-evolves with an optimism bias that may lead to the selection of new equilibria, closely associated with average behavior observed in experimental data. This work reveals both a novel perspective on the evolution of bounded rationality and a co-evolutionary link between the evolution of Theory of Mind and the emergence of misbeliefs.},
note = {DOI: 10.1016/j.isci.2024.108862},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Even though the Theory of Mind in upper primates has been under investigation for decades, how it may evolve remains an open problem. We propose here an evolutionary game theoretical model where a finite population of individuals may use reasoning strategies to infer a response to the anticipated behavior of others within the context of a sequential dilemma, i.e., the Centipede Game. We show that strategies with bounded reasoning evolve and flourish under natural selection, provided they are allowed to make reasoning mistakes and a temptation for higher future gains is in place. We further show that non-deterministic reasoning co-evolves with an optimism bias that may lead to the selection of new equilibria, closely associated with average behavior observed in experimental data. This work reveals both a novel perspective on the evolution of bounded rationality and a co-evolutionary link between the evolution of Theory of Mind and the emergence of misbeliefs. |
Terrucha, Ines; Domingos, Elias Fernández; Santos, Francisco C.; Simoens, Pieter; Lenaerts, Tom The art of compensation: How hybrid teams solve collective-risk dilemmas Journal Article In: PloS one, vol. 19, no. 2 February, 2024, (DOI: 10.1371/journal.pone.0297213). @article{info:hdl:2013/371876b,
title = {The art of compensation: How hybrid teams solve collective-risk dilemmas},
author = {Ines Terrucha and Elias Fernández Domingos and Francisco C. Santos and Pieter Simoens and Tom Lenaerts},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/371876/1/doi_355520.pdf},
year = {2024},
date = {2024-01-01},
journal = {PloS one},
volume = {19},
number = {2 February},
abstract = {It is widely known how the human ability to cooperate has influenced the thriving of our species. However, as we move towards a hybrid human-machine future, it is still unclear how the introduction of artificial agents in our social interactions affect this cooperative capacity. In a one-shot collective risk dilemma, where enough members of a group must cooperate in order to avoid a collective disaster, we study the evolutionary dynamics of cooperation in a hybrid population. In our model, we consider a hybrid population composed of both adaptive and fixed behavior agents. The latter serve as proxies for the machine-like behavior of artificially intelligent agents who implement stochastic strategies previously learned offline. We observe that the adaptive individuals adjust their behavior in function of the presence of artificial agents in their groups to compensate their cooperative (or lack of thereof) efforts. We also find that risk plays a determinant role when assessing whether or not we should form hybrid teams to tackle a collective risk dilemma. When the risk of collective disaster is high, cooperation in the adaptive population falls dramatically in the presence of cooperative artificial agents. A story of compensation, rather than cooperation, where adaptive agents have to secure group success when the artificial agents are not cooperative enough, but will rather not cooperate if the others do so. On the contrary, when risk of collective disaster is low, success is highly improved while cooperation levels within the adaptive population remain the same. Artificial agents can improve the collective success of hybrid teams. However, their application requires a true risk assessment of the situation in order to actually benefit the adaptive population (i.e. the humans) in the long-term.},
note = {DOI: 10.1371/journal.pone.0297213},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
It is widely known how the human ability to cooperate has influenced the thriving of our species. However, as we move towards a hybrid human-machine future, it is still unclear how the introduction of artificial agents in our social interactions affect this cooperative capacity. In a one-shot collective risk dilemma, where enough members of a group must cooperate in order to avoid a collective disaster, we study the evolutionary dynamics of cooperation in a hybrid population. In our model, we consider a hybrid population composed of both adaptive and fixed behavior agents. The latter serve as proxies for the machine-like behavior of artificially intelligent agents who implement stochastic strategies previously learned offline. We observe that the adaptive individuals adjust their behavior in function of the presence of artificial agents in their groups to compensate their cooperative (or lack of thereof) efforts. We also find that risk plays a determinant role when assessing whether or not we should form hybrid teams to tackle a collective risk dilemma. When the risk of collective disaster is high, cooperation in the adaptive population falls dramatically in the presence of cooperative artificial agents. A story of compensation, rather than cooperation, where adaptive agents have to secure group success when the artificial agents are not cooperative enough, but will rather not cooperate if the others do so. On the contrary, when risk of collective disaster is low, success is highly improved while cooperation levels within the adaptive population remain the same. Artificial agents can improve the collective success of hybrid teams. However, their application requires a true risk assessment of the situation in order to actually benefit the adaptive population (i.e. the humans) in the long-term. |
Gravel, Barbara; Renaux, Alexandre; Papadimitriou, Sofia; Smits, Guillaume; Nowe, Ann; Lenaerts, Tom Prioritization of oligogenic variant combinations in whole exomes Journal Article In: Bioinformatics, vol. 40, no. 4, 2024, (DOI: 10.1093/bioinformatics/btae184). @article{info:hdl:2013/374647b,
title = {Prioritization of oligogenic variant combinations in whole exomes},
author = {Barbara Gravel and Alexandre Renaux and Sofia Papadimitriou and Guillaume Smits and Ann Nowe and Tom Lenaerts},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/374647/1/doi_358291.pdf},
year = {2024},
date = {2024-01-01},
journal = {Bioinformatics},
volume = {40},
number = {4},
abstract = {Motivation: Whole exome sequencing (WES) has emerged as a powerful tool for genetic research, enabling the collection of a tremendous amount of data about human genetic variation. However, properly identifying which variants are causative of a genetic disease remains an important challenge, often due to the number of variants that need to be screened. Expanding the screening to combinations of variants in two or more genes, as would be required under the oligogenic inheritance model, simply blows this problem out of proportion. Results: We present here the High-throughput oligogenic prioritizer (Hop), a novel prioritization method that uses direct oligogenic information at the variant, gene and gene pair level to detect digenic variant combinations in WES data. This method leverages information from a knowledge graph, together with specialized pathogenicity predictions in order to effectively rank variant combinations based on how likely they are to explain the patient’s phenotype. The performance of Hop is evaluated in cross-validation on 36 120 synthetic exomes for training and 14 280 additional synthetic exomes for independent testing. Whereas the known pathogenic variant combinations are found in the top 20 in approximately 60% of the cross-validation exomes, 71% are found in the same ranking range when considering the independent set. These results provide a significant improvement over alternative approaches that depend simply on a monogenic assessment of pathogenicity, including early attempts for digenic ranking using monogenic pathogenicity scores.},
note = {DOI: 10.1093/bioinformatics/btae184},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Motivation: Whole exome sequencing (WES) has emerged as a powerful tool for genetic research, enabling the collection of a tremendous amount of data about human genetic variation. However, properly identifying which variants are causative of a genetic disease remains an important challenge, often due to the number of variants that need to be screened. Expanding the screening to combinations of variants in two or more genes, as would be required under the oligogenic inheritance model, simply blows this problem out of proportion. Results: We present here the High-throughput oligogenic prioritizer (Hop), a novel prioritization method that uses direct oligogenic information at the variant, gene and gene pair level to detect digenic variant combinations in WES data. This method leverages information from a knowledge graph, together with specialized pathogenicity predictions in order to effectively rank variant combinations based on how likely they are to explain the patient’s phenotype. The performance of Hop is evaluated in cross-validation on 36 120 synthetic exomes for training and 14 280 additional synthetic exomes for independent testing. Whereas the known pathogenic variant combinations are found in the top 20 in approximately 60% of the cross-validation exomes, 71% are found in the same ranking range when considering the independent set. These results provide a significant improvement over alternative approaches that depend simply on a monogenic assessment of pathogenicity, including early attempts for digenic ranking using monogenic pathogenicity scores. |
Lillepea, Kristiina; Juchnewitsch, Anna Grete; Kasak, Laura; Valkna, Anu; Dutta, Avirup; Pomm, Kristjan; Poolamets, Olev; Nagirnaja, Liina; Tamp, Erik; Mahyari, Eisa; Vihljajev, Vladimir; Tjagur, Stanislav; Papadimitriou, Sofia; Riera-Escamilla, Antoni; Versbraegen, Nassim; Farnetani, Ginevra; Castillo-Madeen, Helen; Sütt, Mailis; Kübarsepp, Viljo; Tennisberg, Sven; Korrovits, Paul; Krausz, Csilla; Aston, Kenneth Ivan; Lenaerts, Tom; Conrad, Donald D. F.; Punab, Margus; Laan, Maris Toward clinical exomes in diagnostics and management of male infertility Journal Article In: American journal of human genetics, vol. 111, no. 5, pp. 877-895, 2024, (DOI: 10.1016/j.ajhg.2024.03.013). @article{info:hdl:2013/374767b,
title = {Toward clinical exomes in diagnostics and management of male infertility},
author = {Kristiina Lillepea and Anna Grete Juchnewitsch and Laura Kasak and Anu Valkna and Avirup Dutta and Kristjan Pomm and Olev Poolamets and Liina Nagirnaja and Erik Tamp and Eisa Mahyari and Vladimir Vihljajev and Stanislav Tjagur and Sofia Papadimitriou and Antoni Riera-Escamilla and Nassim Versbraegen and Ginevra Farnetani and Helen Castillo-Madeen and Mailis Sütt and Viljo Kübarsepp and Sven Tennisberg and Paul Korrovits and Csilla Krausz and Kenneth Ivan Aston and Tom Lenaerts and Donald D. F. Conrad and Margus Punab and Maris Laan},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/374767/3/Lillepeaetal.pdf},
year = {2024},
date = {2024-01-01},
journal = {American journal of human genetics},
volume = {111},
number = {5},
pages = {877-895},
abstract = {Infertility, affecting ∼10% of men, is predominantly caused by primary spermatogenic failure (SPGF). We screened likely pathogenic and pathogenic (LP/P) variants in 638 candidate genes for male infertility in 521 individuals presenting idiopathic SPGF and 323 normozoospermic men in the ESTAND cohort. Molecular diagnosis was reached for 64 men with SPGF (12%), with findings in 39 genes (6%). The yield did not differ significantly between the subgroups with azoospermia (20/185, 11%), oligozoospermia (18/181, 10%), and primary cryptorchidism with SPGF (26/155, 17%). Notably, 19 of 64 LP/P variants (30%) identified in 28 subjects represented recurrent findings in this study and/or with other male infertility cohorts. NR5A1 was the most frequently affected gene, with seven LP/P variants in six SPGF-affected men and two normozoospermic men. The link to SPGF was validated for recently proposed candidate genes ACTRT1, ASZ1, GLUD2, GREB1L, LEO1, RBM5, ROS1, and TGIF2LY. Heterozygous truncating variants in BNC1, reported in female infertility, emerged as plausible causes of severe oligozoospermia. Data suggested that several infertile men may present congenital conditions with less pronounced or pleiotropic phenotypes affecting the development and function of the reproductive system. Genes regulating the hypothalamic-pituitary-gonadal axis were affected in >30% of subjects with LP/P variants. Six individuals had more than one LP/P variant, including five with two findings from the gene panel. A 4-fold increased prevalence of cancer was observed in men with genetic infertility compared to the general male population (8% vs. 2%; p = 4.4 ?x 10−3). Expanding genetic testing in andrology will contribute to the multidisciplinary management of SPGF.},
note = {DOI: 10.1016/j.ajhg.2024.03.013},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Infertility, affecting ∼10% of men, is predominantly caused by primary spermatogenic failure (SPGF). We screened likely pathogenic and pathogenic (LP/P) variants in 638 candidate genes for male infertility in 521 individuals presenting idiopathic SPGF and 323 normozoospermic men in the ESTAND cohort. Molecular diagnosis was reached for 64 men with SPGF (12%), with findings in 39 genes (6%). The yield did not differ significantly between the subgroups with azoospermia (20/185, 11%), oligozoospermia (18/181, 10%), and primary cryptorchidism with SPGF (26/155, 17%). Notably, 19 of 64 LP/P variants (30%) identified in 28 subjects represented recurrent findings in this study and/or with other male infertility cohorts. NR5A1 was the most frequently affected gene, with seven LP/P variants in six SPGF-affected men and two normozoospermic men. The link to SPGF was validated for recently proposed candidate genes ACTRT1, ASZ1, GLUD2, GREB1L, LEO1, RBM5, ROS1, and TGIF2LY. Heterozygous truncating variants in BNC1, reported in female infertility, emerged as plausible causes of severe oligozoospermia. Data suggested that several infertile men may present congenital conditions with less pronounced or pleiotropic phenotypes affecting the development and function of the reproductive system. Genes regulating the hypothalamic-pituitary-gonadal axis were affected in >30% of subjects with LP/P variants. Six individuals had more than one LP/P variant, including five with two findings from the gene panel. A 4-fold increased prevalence of cancer was observed in men with genetic infertility compared to the general male population (8% vs. 2%; p = 4.4 ?x 10−3). Expanding genetic testing in andrology will contribute to the multidisciplinary management of SPGF. |
Nachtegael, Charlotte; Stefani, Jacopo De; Cnudde, Anthony; Lenaerts, Tom DUVEL: an active-learning annotated biomedical corpus for the recognition of oligogenic combinations Journal Article In: Database, vol. 2024, no. 2024, 2024, (DOI: 10.1093/database/baae039). @article{info:hdl:2013/374632b,
title = {DUVEL: an active-learning annotated biomedical corpus for the recognition of oligogenic combinations},
author = {Charlotte Nachtegael and Jacopo De Stefani and Anthony Cnudde and Tom Lenaerts},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/374632/1/doi_358276.pdf},
year = {2024},
date = {2024-01-01},
journal = {Database},
volume = {2024},
number = {2024},
abstract = {Abstract While biomedical relation extraction (bioRE) datasets have been instrumental in the development of methods to support biocuration of single variants from texts, no datasets are currently available for the extraction of digenic or even oligogenic variant relations, despite the reports in literature that epistatic effects between combinations of variants in different loci (or genes) are important to understand disease etiologies. This work presents the creation of a unique dataset of oligogenic variant combinations, geared to train tools to help in the curation of scientific literature. To overcome the hurdles associated with the number of unlabelled instances and the cost of expertise, active learning (AL) was used to optimize the annotation, thus getting assistance in finding the most informative subset of samples to label. By pre-annotating 85 full-text articles containing the relevant relations from the Oligogenic Diseases Database (OLIDA) with PubTator, text fragments featuring potential digenic variant combinations, i.e. gene–variant–gene–variant, were extracted. The resulting fragments of texts were annotated with ALAMBIC, an AL-based annotation platform. The resulting dataset, called DUVEL, is used to fine-tune four state-of-the-art biomedical language models: BiomedBERT, BiomedBERT-large, BioLinkBERT and BioM-BERT. More than 500 000 text fragments were considered for annotation, finally resulting in a dataset with 8442 fragments, 794 of them being positive instances, covering 95% of the original annotated articles. When applied to gene–variant pair detection, BiomedBERT-large achieves the highest F1 score (0.84) after fine-tuning, demonstrating significant improvement compared to the non-fine-tuned model, underlining the relevance of the DUVEL dataset. This study shows how AL may play an important role in the creation of bioRE dataset relevant for biomedical curation applications. DUVEL provides a unique biomedical corpus focusing on 4-ary relations between two genes and two variants. It is made freely available for research on GitHub and Hugging Face. Database URL: https://huggingface.co/datasets/cnachteg/duvel or https://doi.org/10.57967/hf/1571},
note = {DOI: 10.1093/database/baae039},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Abstract While biomedical relation extraction (bioRE) datasets have been instrumental in the development of methods to support biocuration of single variants from texts, no datasets are currently available for the extraction of digenic or even oligogenic variant relations, despite the reports in literature that epistatic effects between combinations of variants in different loci (or genes) are important to understand disease etiologies. This work presents the creation of a unique dataset of oligogenic variant combinations, geared to train tools to help in the curation of scientific literature. To overcome the hurdles associated with the number of unlabelled instances and the cost of expertise, active learning (AL) was used to optimize the annotation, thus getting assistance in finding the most informative subset of samples to label. By pre-annotating 85 full-text articles containing the relevant relations from the Oligogenic Diseases Database (OLIDA) with PubTator, text fragments featuring potential digenic variant combinations, i.e. gene–variant–gene–variant, were extracted. The resulting fragments of texts were annotated with ALAMBIC, an AL-based annotation platform. The resulting dataset, called DUVEL, is used to fine-tune four state-of-the-art biomedical language models: BiomedBERT, BiomedBERT-large, BioLinkBERT and BioM-BERT. More than 500 000 text fragments were considered for annotation, finally resulting in a dataset with 8442 fragments, 794 of them being positive instances, covering 95% of the original annotated articles. When applied to gene–variant pair detection, BiomedBERT-large achieves the highest F1 score (0.84) after fine-tuning, demonstrating significant improvement compared to the non-fine-tuned model, underlining the relevance of the DUVEL dataset. This study shows how AL may play an important role in the creation of bioRE dataset relevant for biomedical curation applications. DUVEL provides a unique biomedical corpus focusing on 4-ary relations between two genes and two variants. It is made freely available for research on GitHub and Hugging Face. Database URL: https://huggingface.co/datasets/cnachteg/duvel or https://doi.org/10.57967/hf/1571 |
Kirchsteiger, Georg; Lenaerts, Tom; Suchon, Remi Voluntary versus mandatory information disclosure in the sequential prisoner’s dilemma Journal Article In: Economic theory, 2024, (DOI: 10.1007/s00199-024-01563-y). @article{info:hdl:2013/373750b,
title = {Voluntary versus mandatory information disclosure in the sequential prisoner’s dilemma},
author = {Georg Kirchsteiger and Tom Lenaerts and Remi Suchon},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/373750/3/s00199-024-01563-y.pdf},
year = {2024},
date = {2024-01-01},
journal = {Economic theory},
abstract = {In sequential social dilemmas with stranger matching, initiating cooperation is inherently risky for the first mover. The disclosure of the second mover’s past actions may be necessary to instigate cooperation. We experimentally compare the effect of mandatory and voluntary disclosure with non-disclosure in a sequential prisoner’s dilemma situation. Our results confirm the positive effects of disclosure on cooperation. We also find that voluntary disclosure is as effective as mandatory disclosure, which runs counter to the results of existing literature on this topic. With voluntary disclosure, second movers who have a good track record chose to disclose, suggesting that they anticipate non-disclosure would signal non-cooperativeness. First movers interpret non-disclosure correctly as a signal of non-cooperativeness. Therefore, they cooperate less than half as often when the second mover decides not to disclose.},
note = {DOI: 10.1007/s00199-024-01563-y},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
In sequential social dilemmas with stranger matching, initiating cooperation is inherently risky for the first mover. The disclosure of the second mover’s past actions may be necessary to instigate cooperation. We experimentally compare the effect of mandatory and voluntary disclosure with non-disclosure in a sequential prisoner’s dilemma situation. Our results confirm the positive effects of disclosure on cooperation. We also find that voluntary disclosure is as effective as mandatory disclosure, which runs counter to the results of existing literature on this topic. With voluntary disclosure, second movers who have a good track record chose to disclose, suggesting that they anticipate non-disclosure would signal non-cooperativeness. First movers interpret non-disclosure correctly as a signal of non-cooperativeness. Therefore, they cooperate less than half as often when the second mover decides not to disclose. |
Terrucha, Ines; Domingos, Elias Fernández; Simoens, Pieter; Lenaerts, Tom Committing to the wrong artificial delegate in a collective-risk dilemma is better than directly committing mistakes Journal Article In: Scientific reports, vol. 14, no. 1, 2024, (DOI: 10.1038/s41598-024-61153-9). @article{info:hdl:2013/374814b,
title = {Committing to the wrong artificial delegate in a collective-risk dilemma is better than directly committing mistakes},
author = {Ines Terrucha and Elias Fernández Domingos and Pieter Simoens and Tom Lenaerts},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/374814/1/doi_358458.pdf},
year = {2024},
date = {2024-01-01},
journal = {Scientific reports},
volume = {14},
number = {1},
abstract = {While autonomous artificial agents are assumed to perfectly execute the strategies they are programmed with, humans who design them may make mistakes. These mistakes may lead to a misalignment between the humans’ intended goals and their agents’ observed behavior, a problem of value alignment. Such an alignment problem may have particularly strong consequences when these autonomous systems are used in social contexts that involve some form of collective risk. By means of an evolutionary game theoretical model, we investigate whether errors in the configuration of artificial agents change the outcome of a collective-risk dilemma, in comparison to a scenario with no delegation. Delegation is here distinguished from no-delegation simply by the moment at which a mistake occurs: either when programming/choosing the agent (in case of delegation) or when executing the actions at each round of the game (in case of no-delegation). We find that, while errors decrease success rate, it is better to delegate and commit to a somewhat flawed strategy, perfectly executed by an autonomous agent, than to commit execution errors directly. Our model also shows that in the long-term, delegation strategies should be favored over no-delegation, if given the choice.},
note = {DOI: 10.1038/s41598-024-61153-9},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
While autonomous artificial agents are assumed to perfectly execute the strategies they are programmed with, humans who design them may make mistakes. These mistakes may lead to a misalignment between the humans’ intended goals and their agents’ observed behavior, a problem of value alignment. Such an alignment problem may have particularly strong consequences when these autonomous systems are used in social contexts that involve some form of collective risk. By means of an evolutionary game theoretical model, we investigate whether errors in the configuration of artificial agents change the outcome of a collective-risk dilemma, in comparison to a scenario with no delegation. Delegation is here distinguished from no-delegation simply by the moment at which a mistake occurs: either when programming/choosing the agent (in case of delegation) or when executing the actions at each round of the game (in case of no-delegation). We find that, while errors decrease success rate, it is better to delegate and commit to a somewhat flawed strategy, perfectly executed by an autonomous agent, than to commit execution errors directly. Our model also shows that in the long-term, delegation strategies should be favored over no-delegation, if given the choice. |
Rivière, Quentin; Raskin, Virginie; Melo, Romário; Boutet, Stéphanie; Corso, Massimiliano; Defrance, Matthieu; Webb, Alex A. R.; Verbruggen, Nathalie; Anoman, Djoro Armand Effects of light regimes on circadian gene co‐expression networks in Arabidopsis thaliana Journal Article In: Plant Direct, vol. 8, no. 8, 2024, (DOI: 10.1002/pld3.70001). @article{info:hdl:2013/384388b,
title = {Effects of light regimes on circadian gene co‐expression networks in Arabidopsis thaliana},
author = {Quentin Rivière and Virginie Raskin and Romário Melo and Stéphanie Boutet and Massimiliano Corso and Matthieu Defrance and Alex A. R. Webb and Nathalie Verbruggen and Djoro Armand Anoman},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/384388/3/Plant-Direct-2024.pdf},
year = {2024},
date = {2024-01-01},
journal = {Plant Direct},
volume = {8},
number = {8},
abstract = {Abstract Light/dark (LD) cycles are responsible for oscillations in gene expression, which modulate several aspects of plant physiology. Those oscillations can persist under constant conditions due to regulation by the circadian oscillator. The response of the transcriptome to light regimes is dynamic and allows plants to adapt rapidly to changing environmental conditions. We compared the transcriptome of Arabidopsis under LD and constant light (LL) for 3 days and identified different gene co‐expression networks in the two light regimes. Our studies yielded unforeseen insights into circadian regulation. Intuitively, we anticipated that gene clusters regulated by the circadian oscillator would display oscillations under LD cycles. However, we found transcripts encoding components of the flavonoid metabolism pathway that were rhythmic in LL but not in LD. We also discovered that the expressions of many stress‐related genes were significantly increased during the dark period in LD relative to the subjective night in LL, whereas the expression of these genes in the light period was similar. The nocturnal pattern of these stress‐related gene expressions suggested a form of “skotoprotection.” The transcriptomics data were made available in a web application named Cyclath , which we believe will be a useful tool to contribute to a better understanding of the impact of light regimes on plants.},
note = {DOI: 10.1002/pld3.70001},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Abstract Light/dark (LD) cycles are responsible for oscillations in gene expression, which modulate several aspects of plant physiology. Those oscillations can persist under constant conditions due to regulation by the circadian oscillator. The response of the transcriptome to light regimes is dynamic and allows plants to adapt rapidly to changing environmental conditions. We compared the transcriptome of Arabidopsis under LD and constant light (LL) for 3 days and identified different gene co‐expression networks in the two light regimes. Our studies yielded unforeseen insights into circadian regulation. Intuitively, we anticipated that gene clusters regulated by the circadian oscillator would display oscillations under LD cycles. However, we found transcripts encoding components of the flavonoid metabolism pathway that were rhythmic in LL but not in LD. We also discovered that the expressions of many stress‐related genes were significantly increased during the dark period in LD relative to the subjective night in LL, whereas the expression of these genes in the light period was similar. The nocturnal pattern of these stress‐related gene expressions suggested a form of “skotoprotection.” The transcriptomics data were made available in a web application named Cyclath , which we believe will be a useful tool to contribute to a better understanding of the impact of light regimes on plants. |
Jansen, Maarten Information criteria for structured parameter selection in high dimensional tree and graph models Journal Article In: Digital signal processing, vol. 148, 2024, (Language of publication: fr). @article{info:hdl:2013/372845b,
title = {Information criteria for structured parameter selection in high dimensional tree and graph models},
author = {Maarten Jansen},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/372845/3/jansen24structuredpreprint.pdf},
year = {2024},
date = {2024-01-01},
journal = {Digital signal processing},
volume = {148},
note = {Language of publication: fr},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
|
Bhattacharya, Shreya; Lefèvre, Laure; Chatzistergos, T; Hayakawa, Hisashi; Jansen, Maarten RudolfWolf to AlfredWolfer: The Transfer of the Reference Observer in the International Sunspot Number Series (1876–1893) Journal Article In: Solar physics, vol. 299, 2024, (Language of publication: fr). @article{info:hdl:2013/372844b,
title = {RudolfWolf to AlfredWolfer: The Transfer of the Reference Observer in the International Sunspot Number Series (1876–1893)},
author = {Shreya Bhattacharya and Laure Lefèvre and T Chatzistergos and Hisashi Hayakawa and Maarten Jansen},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/372844/3/PreprintBhattacharya202404.pdf},
year = {2024},
date = {2024-01-01},
journal = {Solar physics},
volume = {299},
note = {Language of publication: fr},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
|
Colot, Martin; Simar, Cédric; Petieau, Mathieu; Alvarez, Ana Maria Cebolla; Chéron, Guy; Bontempi, Gianluca EMG subspace alignment and visualization for cross-subject hand gesture classification Miscellaneous 2024, (Conference: ECML-PKDD 2023 Worshop - Adapting to change : Reliable Learning Across Domains (2023-09-18: Turin)). @misc{info:hdl:2013/373864b,
title = {EMG subspace alignment and visualization for cross-subject hand gesture classification},
author = {Martin Colot and Cédric Simar and Mathieu Petieau and Ana Maria Cebolla Alvarez and Guy Chéron and Gianluca Bontempi},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/373864/3/ECML_PKDD_2023__Martin_Colot_Main_and_Appendix.pdf},
year = {2024},
date = {2024-01-01},
abstract = {Electromyograms (EMG)-based hand gesture recognition systems are a promising technology for human/machine interfaces. However, one of their main limitations is the long calibration time that is typically required to handle new users. The paper discusses and analyses the challenge of cross-subject generalization thanks to an original dataset containing the EMG signals of 14 human subjects during hand gestures. The experimental results show that, though an accurate generalization based on pooling multiple subjects is hardly achievable, it is possible to improve the cross-subject estimation by identifying a robust low-dimensional subspace for multiple subjects and aligning it to a target subject. A visualization of the subspace enables us to provide insights for the improvement of cross-subject generalization with EMG signals.},
note = {Conference: ECML-PKDD 2023 Worshop - Adapting to change : Reliable Learning Across Domains (2023-09-18: Turin)},
keywords = {},
pubstate = {published},
tppubtype = {misc}
}
Electromyograms (EMG)-based hand gesture recognition systems are a promising technology for human/machine interfaces. However, one of their main limitations is the long calibration time that is typically required to handle new users. The paper discusses and analyses the challenge of cross-subject generalization thanks to an original dataset containing the EMG signals of 14 human subjects during hand gestures. The experimental results show that, though an accurate generalization based on pooling multiple subjects is hardly achievable, it is possible to improve the cross-subject estimation by identifying a robust low-dimensional subspace for multiple subjects and aligning it to a target subject. A visualization of the subspace enables us to provide insights for the improvement of cross-subject generalization with EMG signals. |
Lebichot, Bertrand; Siblini, Wissam; Paldino, Gian Marco; Borgne, Yann-Aël Le; Oblé, Frédéric; Bontempi, Gianluca Assessment of catastrophic forgetting in continual credit card fraud detection Journal Article In: Expert systems with applications, vol. 249, 2024, (DOI: 10.1016/j.eswa.2024.123445). @article{info:hdl:2013/370795b,
title = {Assessment of catastrophic forgetting in continual credit card fraud detection},
author = {Bertrand Lebichot and Wissam Siblini and Gian Marco Paldino and Yann-Aël Le Borgne and Frédéric Oblé and Gianluca Bontempi},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/370795/3/preproofs-main.pdf},
year = {2024},
date = {2024-01-01},
journal = {Expert systems with applications},
volume = {249},
abstract = {The volume of e-commerce continues to increase year after year. Buying goods on the internet is easy and practical, and took a huge boost during the lockdowns of the Covid crisis. However, this is also an open window for fraudsters and the corresponding financial loss costs billions of dollars. In this paper, we study e-commerce credit card fraud detection, in collaboration with our industrial partner, Worldline. Transactional companies are more and more dependent on machine learning models such as deep learning anomaly detection models, as part of real-world fraud detection systems (FDS). We focus on continual learning to find the best model with respect to two objectives: to maximize the accuracy and to minimize the catastrophic forgetting phenomenon. For the latter, we proposed an evaluation procedure to quantify the forgetting in data streams with delayed feedback: the plasticity/stability visualization matrix. We also investigated six strategies and 13 methods on a real-size case study including five months of e-commerce credit card transactions. Finally, we discuss how the trade-off between plasticity and stability is set, in practice, in the case of FDS.},
note = {DOI: 10.1016/j.eswa.2024.123445},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
The volume of e-commerce continues to increase year after year. Buying goods on the internet is easy and practical, and took a huge boost during the lockdowns of the Covid crisis. However, this is also an open window for fraudsters and the corresponding financial loss costs billions of dollars. In this paper, we study e-commerce credit card fraud detection, in collaboration with our industrial partner, Worldline. Transactional companies are more and more dependent on machine learning models such as deep learning anomaly detection models, as part of real-world fraud detection systems (FDS). We focus on continual learning to find the best model with respect to two objectives: to maximize the accuracy and to minimize the catastrophic forgetting phenomenon. For the latter, we proposed an evaluation procedure to quantify the forgetting in data streams with delayed feedback: the plasticity/stability visualization matrix. We also investigated six strategies and 13 methods on a real-size case study including five months of e-commerce credit card transactions. Finally, we discuss how the trade-off between plasticity and stability is set, in practice, in the case of FDS. |
Paldino, Gian Marco; Lebichot, Bertrand; Borgne, Yann-Aël Le; Siblini, Wissam; Oblé, Frédéric; Boracchi, Giacomo; Bontempi, Gianluca The role of diversity and ensemble learning in credit card fraud detection Journal Article In: Advances in Data Analysis and Classification, vol. 18, no. 1, pp. 193-217, 2024, (DOI: 10.1007/s11634-022-00515-5). @article{info:hdl:2013/372242b,
title = {The role of diversity and ensemble learning in credit card fraud detection},
author = {Gian Marco Paldino and Bertrand Lebichot and Yann-Aël Le Borgne and Wissam Siblini and Frédéric Oblé and Giacomo Boracchi and Gianluca Bontempi},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/372242/1/elsevier_355886.pdf},
year = {2024},
date = {2024-01-01},
journal = {Advances in Data Analysis and Classification},
volume = {18},
number = {1},
pages = {193-217},
abstract = {The number of daily credit card transactions is inexorably growing: the e-commerce market expansion and the recent constraints for the Covid-19 pandemic have significantly increased the use of electronic payments. The ability to precisely detect fraudulent transactions is increasingly important, and machine learning models are now a key component of the detection process. Standard machine learning techniques are widely employed, but inadequate for the evolving nature of customers behavior entailing continuous changes in the underlying data distribution. his problem is often tackled by discarding past knowledge, despite its potential relevance in the case of recurrent concepts. Appropriate exploitation of historical knowledge is necessary: we propose a learning strategy that relies on diversity-based ensemble learning and allows to preserve past concepts and reuse them for a faster adaptation to changes. In our experiments, we adopt several state-of-the-art diversity measures and we perform comparisons with various other learning approaches. We assess the effectiveness of our proposed learning strategy on extracts of two real datasets from two European countries, containing more than 30 M and 50 M transactions, provided by our industrial partner, Worldline, a leading company in the field.},
note = {DOI: 10.1007/s11634-022-00515-5},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
The number of daily credit card transactions is inexorably growing: the e-commerce market expansion and the recent constraints for the Covid-19 pandemic have significantly increased the use of electronic payments. The ability to precisely detect fraudulent transactions is increasingly important, and machine learning models are now a key component of the detection process. Standard machine learning techniques are widely employed, but inadequate for the evolving nature of customers behavior entailing continuous changes in the underlying data distribution. his problem is often tackled by discarding past knowledge, despite its potential relevance in the case of recurrent concepts. Appropriate exploitation of historical knowledge is necessary: we propose a learning strategy that relies on diversity-based ensemble learning and allows to preserve past concepts and reuse them for a faster adaptation to changes. In our experiments, we adopt several state-of-the-art diversity measures and we perform comparisons with various other learning approaches. We assess the effectiveness of our proposed learning strategy on extracts of two real datasets from two European countries, containing more than 30 M and 50 M transactions, provided by our industrial partner, Worldline, a leading company in the field. |
Simar, Cédric; Colot, Martin; Alvarez, Ana Maria Cebolla; Petieau, Mathieu; Chéron, Guy; Bontempi, Gianluca Machine learning for hand pose classification from phasic and tonic EMG signals during bimanual activities in virtual reality Journal Article In: Frontiers in Neuroscience, vol. 18, 2024, (DOI: 10.3389/fnins.2024.1329411). @article{info:hdl:2013/373455b,
title = {Machine learning for hand pose classification from phasic and tonic EMG signals during bimanual activities in virtual reality},
author = {Cédric Simar and Martin Colot and Ana Maria Cebolla Alvarez and Mathieu Petieau and Guy Chéron and Gianluca Bontempi},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/373455/1/doi_357099.pdf},
year = {2024},
date = {2024-01-01},
journal = {Frontiers in Neuroscience},
volume = {18},
abstract = {Myoelectric prostheses have recently shown significant promise for restoring hand function in individuals with upper limb loss or deficiencies, driven by advances in machine learning and increasingly accessible bioelectrical signal acquisition devices. Here, we first introduce and validate a novel experimental paradigm using a virtual reality headset equipped with hand-tracking capabilities to facilitate the recordings of synchronized EMG signals and hand pose estimation. Using both the phasic and tonic EMG components of data acquired through the proposed paradigm, we compare hand gesture classification pipelines based on standard signal processing features, convolutional neural networks, and covariance matrices with Riemannian geometry computed from raw or xDAWN-filtered EMG signals. We demonstrate the performance of the latter for gesture classification using EMG signals. We further hypothesize that introducing physiological knowledge in machine learning models will enhance their performances, leading to better myoelectric prosthesis control. We demonstrate the potential of this approach by using the neurophysiological integration of the “move command" to better separate the phasic and tonic components of the EMG signals, significantly improving the performance of sustained posture recognition. These results pave the way for the development of new cutting-edge machine learning techniques, likely refined by neurophysiology, that will further improve the decoding of real-time natural gestures and, ultimately, the control of myoelectric prostheses.},
note = {DOI: 10.3389/fnins.2024.1329411},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Myoelectric prostheses have recently shown significant promise for restoring hand function in individuals with upper limb loss or deficiencies, driven by advances in machine learning and increasingly accessible bioelectrical signal acquisition devices. Here, we first introduce and validate a novel experimental paradigm using a virtual reality headset equipped with hand-tracking capabilities to facilitate the recordings of synchronized EMG signals and hand pose estimation. Using both the phasic and tonic EMG components of data acquired through the proposed paradigm, we compare hand gesture classification pipelines based on standard signal processing features, convolutional neural networks, and covariance matrices with Riemannian geometry computed from raw or xDAWN-filtered EMG signals. We demonstrate the performance of the latter for gesture classification using EMG signals. We further hypothesize that introducing physiological knowledge in machine learning models will enhance their performances, leading to better myoelectric prosthesis control. We demonstrate the potential of this approach by using the neurophysiological integration of the “move command" to better separate the phasic and tonic components of the EMG signals, significantly improving the performance of sustained posture recognition. These results pave the way for the development of new cutting-edge machine learning techniques, likely refined by neurophysiology, that will further improve the decoding of real-time natural gestures and, ultimately, the control of myoelectric prostheses. |
Cerqueira, Vitor; Torgo, Luis; Bontempi, Gianluca Instance-based meta-learning for conditionally dependent univariate multi-step forecasting Journal Article In: International journal of forecasting, 2024, (DOI: 10.1016/j.ijforecast.2023.12.010). @article{info:hdl:2013/371938b,
title = {Instance-based meta-learning for conditionally dependent univariate multi-step forecasting},
author = {Vitor Cerqueira and Luis Torgo and Gianluca Bontempi},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/371938/3/ijf.meta2.pdf},
year = {2024},
date = {2024-01-01},
journal = {International journal of forecasting},
abstract = {Multi-step prediction is a key challenge in univariate forecasting. However, forecasting accuracy decreases as predictions are made further into the future. This is caused by the decreasing predictability and the error propagation along the horizon. In this paper, we propose a novel method called Forecasted Trajectory Neighbors (FTN) for multi-step forecasting with univariate time series. FTN is a meta-learning strategy that can be integrated with any state-of-the-art multi-step forecasting approach. It works by using training observations to correct the errors made during multiple predictions. This is accomplished by retrieving the nearest neighbors of the multi-step forecasts and averaging these for prediction. The motivation is to introduce, in a lightweight manner, a conditional dependent constraint across the forecasting horizons. Such a constraint, not always taken into account by most strategies, can be considered as a sort of regularization element. We carried out extensive experiments using 7795 time series from different application domains. We found that our method improves the performance of several state-of-the-art multi-step forecasting methods. An implementation of the proposed method is publicly available online, and the experiments are reproducible.},
note = {DOI: 10.1016/j.ijforecast.2023.12.010},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Multi-step prediction is a key challenge in univariate forecasting. However, forecasting accuracy decreases as predictions are made further into the future. This is caused by the decreasing predictability and the error propagation along the horizon. In this paper, we propose a novel method called Forecasted Trajectory Neighbors (FTN) for multi-step forecasting with univariate time series. FTN is a meta-learning strategy that can be integrated with any state-of-the-art multi-step forecasting approach. It works by using training observations to correct the errors made during multiple predictions. This is accomplished by retrieving the nearest neighbors of the multi-step forecasts and averaging these for prediction. The motivation is to introduce, in a lightweight manner, a conditional dependent constraint across the forecasting horizons. Such a constraint, not always taken into account by most strategies, can be considered as a sort of regularization element. We carried out extensive experiments using 7795 time series from different application domains. We found that our method improves the performance of several state-of-the-art multi-step forecasting methods. An implementation of the proposed method is publicly available online, and the experiments are reproducible. |
Terrucha, Ines; Domingos, Elias Fernandez; Suchon, Remi; Santos, Francisco C; Simoens, Pieter; Lenaerts, Tom Humans program artificial delegates to accurately solve collective-risk dilemmas, but lack precision Miscellaneous 2024, (Conference: Machine+behavior Conference(Berlin, Allemagne)). @misc{info:hdl:2013/385912,
title = {Humans program artificial delegates to accurately solve collective-risk dilemmas, but lack precision},
author = {Ines Terrucha and Elias Fernandez Domingos and Remi Suchon and Francisco C Santos and Pieter Simoens and Tom Lenaerts},
url = {http://hdl.handle.net/2013/ULB-DIPOT:oai:dipot.ulb.ac.be:2013/385912},
year = {2024},
date = {2024-01-01},
note = {Conference: Machine+behavior Conference(Berlin, Allemagne)},
keywords = {},
pubstate = {published},
tppubtype = {misc}
}
|
Kirchsteiger, Georg; Lenaerts, Tom; Suchon, Remi Growing cooperation Miscellaneous 2024, (Conference: Conference of the French Experimental Economics Association(14: grenoble, France)). @misc{info:hdl:2013/385911,
title = {Growing cooperation},
author = {Georg Kirchsteiger and Tom Lenaerts and Remi Suchon},
url = {http://hdl.handle.net/2013/ULB-DIPOT:oai:dipot.ulb.ac.be:2013/385911},
year = {2024},
date = {2024-01-01},
note = {Conference: Conference of the French Experimental Economics Association(14: grenoble, France)},
keywords = {},
pubstate = {published},
tppubtype = {misc}
}
|
Bosch, Inas; Gravel, Barbara; Lenaerts, Tom Knowledge graph embeddings for the prediction of pathogenic gene pairs Miscellaneous 2024, (Conference: European Conference on Computational Biology.(23: 16/09-20/09/2024: Turku, Finland)). @misc{info:hdl:2013/385910,
title = {Knowledge graph embeddings for the prediction of pathogenic gene pairs},
author = {Inas Bosch and Barbara Gravel and Tom Lenaerts},
url = {http://hdl.handle.net/2013/ULB-DIPOT:oai:dipot.ulb.ac.be:2013/385910},
year = {2024},
date = {2024-01-01},
note = {Conference: European Conference on Computational Biology.(23: 16/09-20/09/2024: Turku, Finland)},
keywords = {},
pubstate = {published},
tppubtype = {misc}
}
|
Gravel, Barbara; Renaux, Alexandre; Papadimitriou, Sofia; Smits, Guillaume; Nowé, Ann; Lenaerts, Tom Prioritization of variant combinations in whole exomes Miscellaneous 2024, (Conference: European Conference on Computational Biology.(23: 16/09-20/09/2024: Turku, Finland)). @misc{info:hdl:2013/385909,
title = {Prioritization of variant combinations in whole exomes},
author = {Barbara Gravel and Alexandre Renaux and Sofia Papadimitriou and Guillaume Smits and Ann Nowé and Tom Lenaerts},
url = {http://hdl.handle.net/2013/ULB-DIPOT:oai:dipot.ulb.ac.be:2013/385909},
year = {2024},
date = {2024-01-01},
note = {Conference: European Conference on Computational Biology.(23: 16/09-20/09/2024: Turku, Finland)},
keywords = {},
pubstate = {published},
tppubtype = {misc}
}
|
Abels, Axel; Lenaerts, Tom; Trianni, Vito; Nowé, Ann Dealing with Expert Bias in Collective Decision-making Miscellaneous 2024, (Conference: European Conference on Artificial Intelligence(27: 19/10-24/10/2024: Santiago de Compostella)). @misc{info:hdl:2013/385908,
title = {Dealing with Expert Bias in Collective Decision-making},
author = {Axel Abels and Tom Lenaerts and Vito Trianni and Ann Nowé},
url = {http://hdl.handle.net/2013/ULB-DIPOT:oai:dipot.ulb.ac.be:2013/385908},
year = {2024},
date = {2024-01-01},
note = {Conference: European Conference on Artificial Intelligence(27: 19/10-24/10/2024: Santiago de Compostella)},
keywords = {},
pubstate = {published},
tppubtype = {misc}
}
|
Leung, Chin Wing; Lenaerts, Tom; Turrini, Paolo To Promote Full Cooperation in Social Dilemmas, Agents Need to Unlearn Loyalty Proceedings Article In: Larson, Kate (Ed.): Proceedings of the Thirty-Third International Joint Conference on Artificial Intelligence, pp. 111-119, International Joint Conferences on Artificial Intelligence (IJCAI) Organization, 2024, (Conference: International Joint Conference on Artificial Intelligence(33: 3/8-9/8/2024: Jeju. Korea)). @inproceedings{info:hdl:2013/385907,
title = {To Promote Full Cooperation in Social Dilemmas, Agents Need to Unlearn Loyalty},
author = {Chin Wing Leung and Tom Lenaerts and Paolo Turrini},
editor = {Kate Larson},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/385907/3/0013.pdf},
year = {2024},
date = {2024-01-01},
booktitle = {Proceedings of the Thirty-Third International Joint Conference on Artificial Intelligence},
pages = {111-119},
publisher = {International Joint Conferences on Artificial Intelligence (IJCAI) Organization},
abstract = {If given the choice, what strategy should agents use to switch partners in strategic social interactions? While many analyses have been performed on specific switching heuristics, showing how and when these lead to more cooperation, no insights have been provided into which rule will actually be learnt by agents when given the freedom to do so. Starting from a baseline model that has demonstrated the potential of rewiring for cooperation, we provide answers to this question over the full spectrum of social dilemmas. Multi-agent Q-learning with Boltzmann exploration is used to learn when to sever or maintain an association. In both the Prisoner's Dilemma and the Stag Hunt games we observe that the Out-for-Tat rewiring rule, breaking ties with other agents choosing socially undesirable actions, becomes dominant, confirming at the same time that cooperation flourishes when rewiring is fast enough relative to imitation. Nonetheless, in the transitory region before full cooperation, a Stay strategy, keeping a connection at all costs, remains present, which shows that loyalty needs to be overcome for full cooperation to emerge. In conclusion, individuals learn cooperation-promoting rewiring rules but need to overcome a kind of loyalty to achieve full cooperation in the full spectrum of social dilemmas.},
note = {Conference: International Joint Conference on Artificial Intelligence(33: 3/8-9/8/2024: Jeju. Korea)},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
If given the choice, what strategy should agents use to switch partners in strategic social interactions? While many analyses have been performed on specific switching heuristics, showing how and when these lead to more cooperation, no insights have been provided into which rule will actually be learnt by agents when given the freedom to do so. Starting from a baseline model that has demonstrated the potential of rewiring for cooperation, we provide answers to this question over the full spectrum of social dilemmas. Multi-agent Q-learning with Boltzmann exploration is used to learn when to sever or maintain an association. In both the Prisoner's Dilemma and the Stag Hunt games we observe that the Out-for-Tat rewiring rule, breaking ties with other agents choosing socially undesirable actions, becomes dominant, confirming at the same time that cooperation flourishes when rewiring is fast enough relative to imitation. Nonetheless, in the transitory region before full cooperation, a Stay strategy, keeping a connection at all costs, remains present, which shows that loyalty needs to be overcome for full cooperation to emerge. In conclusion, individuals learn cooperation-promoting rewiring rules but need to overcome a kind of loyalty to achieve full cooperation in the full spectrum of social dilemmas. |
Molinghen, Yannick; Avalos, Raphaël; Achter, Mark Van; Nowé, Ann; Lenaerts, Tom Laser Learning Environment: A new environment for coordination-critical multi-agent tasks Proceedings Article In: Oliehoek, Frans F. A.; Manon, Kok; Verwer, Sicco (Ed.): Artificial Intelligence and Machine Learning: Revised Selected Papers, Springer Science and Business Media Deutschland GmbH, 2024, (Conference: Benelux Conference Ai conference, BNAIC(35: 8-10/11/2023: TU Delft)). @inproceedings{info:hdl:2013/370546,
title = {Laser Learning Environment: A new environment for coordination-critical multi-agent tasks},
author = {Yannick Molinghen and Raphaël Avalos and Mark Van Achter and Ann Nowé and Tom Lenaerts},
editor = {Frans F. A. Oliehoek and Kok Manon and Sicco Verwer},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/370546/4/2404.03596v1.pdf},
year = {2024},
date = {2024-01-01},
booktitle = {Artificial Intelligence and Machine Learning: Revised Selected Papers},
publisher = {Springer Science and Business Media Deutschland GmbH},
series = {Communications in Computer and Information Science},
abstract = {We introduce the Laser Learning Environment (LLE), a collaborative multi-agent reinforcement learning environment where coordination is key. In LLE, agents depend on each other to make progress (interdependence), must jointly take specific sequences of actions to succeed (perfect coordination), and accomplishing those joint actions does not yield any intermediate reward (zero-incentive dynamics). The challenge of such problems lies in the difficulty of escaping state space bottlenecks caused by interdependence steps since escaping those bottlenecks is not rewarded. We test multiple state-of-the-art value-based MARL algorithms against LLE and show that they consistently fail at the collaborative task because of their inability to escape state space bottlenecks, even though they successfully achieve perfect coordination. We show that Q-learning extensions such as prioritised experience replay and n-steps return hinder exploration in environments with zero-incentive dynamics, and find that intrinsic curiosity with random network distillation is not sufficient to escape those bottlenecks. We demonstrate the need for novel methods to solve this problem and the relevance of LLE as cooperative MARL benchmark.},
note = {Conference: Benelux Conference Ai conference, BNAIC(35: 8-10/11/2023: TU Delft)},
keywords = {},
pubstate = {published},
tppubtype = {inproceedings}
}
We introduce the Laser Learning Environment (LLE), a collaborative multi-agent reinforcement learning environment where coordination is key. In LLE, agents depend on each other to make progress (interdependence), must jointly take specific sequences of actions to succeed (perfect coordination), and accomplishing those joint actions does not yield any intermediate reward (zero-incentive dynamics). The challenge of such problems lies in the difficulty of escaping state space bottlenecks caused by interdependence steps since escaping those bottlenecks is not rewarded. We test multiple state-of-the-art value-based MARL algorithms against LLE and show that they consistently fail at the collaborative task because of their inability to escape state space bottlenecks, even though they successfully achieve perfect coordination. We show that Q-learning extensions such as prioritised experience replay and n-steps return hinder exploration in environments with zero-incentive dynamics, and find that intrinsic curiosity with random network distillation is not sufficient to escape those bottlenecks. We demonstrate the need for novel methods to solve this problem and the relevance of LLE as cooperative MARL benchmark. |
Attafi, Omar Abdelghani; Clementel, Damiano; Kyritsis, Konstantinos; Capriotti, Emidio; Farrell, Gavin; Fragkouli, Styliani-Christina; Castro, Leyla Jael; Hatos, András; Lenaerts, Tom; Mazurenko, Stanislav; Mozaffari, Soroush; Pradelli, Franco; Ruch, Patrick; Savojardo, Castrense; Turina, Maria Paola; Zambelli, Federico; Piovesan, Damiano; Monzon, Alexander Miguel; Psomopoulos, Fotis F. E.; Tosatto, Silvio S. C. E. DOME Registry: implementing community-wide recommendations for reporting supervised machine learning in biology Journal Article In: GigaScience, vol. 13, pp. 8, 2024, (DOI: 10.1093/gigascience/giae094). @article{info:hdl:2013/385906,
title = {DOME Registry: implementing community-wide recommendations for reporting supervised machine learning in biology},
author = {Omar Abdelghani Attafi and Damiano Clementel and Konstantinos Kyritsis and Emidio Capriotti and Gavin Farrell and Styliani-Christina Fragkouli and Leyla Jael Castro and András Hatos and Tom Lenaerts and Stanislav Mazurenko and Soroush Mozaffari and Franco Pradelli and Patrick Ruch and Castrense Savojardo and Maria Paola Turina and Federico Zambelli and Damiano Piovesan and Alexander Miguel Monzon and Fotis F. E. Psomopoulos and Silvio S. C. E. Tosatto},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/385906/3/giae094-2.pdf},
year = {2024},
date = {2024-01-01},
journal = {GigaScience},
volume = {13},
pages = {8},
abstract = {Abstract Supervised machine learning (ML) is used extensively in biology and deserves closer scrutiny. The Data Optimization Model Evaluation (DOME) recommendations aim to enhance the validation and reproducibility of ML research by establishing standards for key aspects such as data handling and processing, optimization, evaluation, and model interpretability. The recommendations help to ensure that key details are reported transparently by providing a structured set of questions. Here, we introduce the DOME registry (URL: registry.dome-ml.org), a database that allows scientists to manage and access comprehensive DOME-related information on published ML studies. The registry uses external resources like ORCID, APICURON, and the Data Stewardship Wizard to streamline the annotation process and ensure comprehensive documentation. By assigning unique identifiers and DOME scores to publications, the registry fosters a standardized evaluation of ML methods. Future plans include continuing to grow the registry through community curation, improving the DOME score definition and encouraging publishers to adopt DOME standards, and promoting transparency and reproducibility of ML in the life sciences.},
note = {DOI: 10.1093/gigascience/giae094},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Abstract Supervised machine learning (ML) is used extensively in biology and deserves closer scrutiny. The Data Optimization Model Evaluation (DOME) recommendations aim to enhance the validation and reproducibility of ML research by establishing standards for key aspects such as data handling and processing, optimization, evaluation, and model interpretability. The recommendations help to ensure that key details are reported transparently by providing a structured set of questions. Here, we introduce the DOME registry (URL: registry.dome-ml.org), a database that allows scientists to manage and access comprehensive DOME-related information on published ML studies. The registry uses external resources like ORCID, APICURON, and the Data Stewardship Wizard to streamline the annotation process and ensure comprehensive documentation. By assigning unique identifiers and DOME scores to publications, the registry fosters a standardized evaluation of ML methods. Future plans include continuing to grow the registry through community curation, improving the DOME score definition and encouraging publishers to adopt DOME standards, and promoting transparency and reproducibility of ML in the life sciences. |
Juchnewitsch, Anna Grete; Pomm, Kristjan; Dutta, Avirup; Tamp, Erik; Valkna, Anu; Lillepea, Kristiina; Mahyari, Eisa; Tjagur, Stanislav; Belova, Galina; Kübarsepp, Viljo; Castillo-Madeen, Helen; Riera-Escamilla, Antoni; Põlluaas, Lisanna; Nagirnaja, Liina; Poolamets, Olev; Vihljajev, Vladimir; Sütt, Mailis; Versbraegen, Nassim; Papadimitriou, Sofia; McLachlan, Robert Ian; Jarvi, Keith Allen; Schlegel, Peter P. N.; Tennisberg, Sven; Korrovits, Paul; Vigh-Conrad, Katinka; O’Bryan, Moira M. K.; Aston, Kenneth Ivan; Lenaerts, Tom; Conrad, Donald D. F.; Kasak, Laura; Punab, Margus; Laan, Maris Undiagnosed RASopathies in infertile men Journal Article In: Frontiers in endocrinology, vol. 15, 2024, (DOI: 10.3389/fendo.2024.1312357). @article{info:hdl:2013/374860,
title = {Undiagnosed RASopathies in infertile men},
author = {Anna Grete Juchnewitsch and Kristjan Pomm and Avirup Dutta and Erik Tamp and Anu Valkna and Kristiina Lillepea and Eisa Mahyari and Stanislav Tjagur and Galina Belova and Viljo Kübarsepp and Helen Castillo-Madeen and Antoni Riera-Escamilla and Lisanna Põlluaas and Liina Nagirnaja and Olev Poolamets and Vladimir Vihljajev and Mailis Sütt and Nassim Versbraegen and Sofia Papadimitriou and Robert Ian McLachlan and Keith Allen Jarvi and Peter P. N. Schlegel and Sven Tennisberg and Paul Korrovits and Katinka Vigh-Conrad and Moira M. K. O’Bryan and Kenneth Ivan Aston and Tom Lenaerts and Donald D. F. Conrad and Laura Kasak and Margus Punab and Maris Laan},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/374860/1/doi_358504.pdf},
year = {2024},
date = {2024-01-01},
journal = {Frontiers in endocrinology},
volume = {15},
abstract = {RASopathies are syndromes caused by congenital defects in the Ras/mitogen-activated protein kinase (MAPK) pathway genes, with a population prevalence of 1 in 1,000. Patients are typically identified in childhood based on diverse characteristic features, including cryptorchidism (CR) in >50% of affected men. As CR predisposes to spermatogenic failure (SPGF; total sperm count per ejaculate 0–39 million), we hypothesized that men seeking infertility management include cases with undiagnosed RASopathies. Likely pathogenic or pathogenic (LP/P) variants in 22 RASopathy-linked genes were screened in 521 idiopathic SPGF patients (including 155 CR cases) and 323 normozoospermic controls using exome sequencing. All 844 men were recruited to the ESTonian ANDrology (ESTAND) cohort and underwent identical andrological phenotyping. RASopathy-specific variant interpretation guidelines were used for pathogenicity assessment. LP/P variants were identified in PTPN11 (two), SOS1 (three), SOS2 (one), LZTR1 (one), SPRED1 (one), NF1 (one), and MAP2K1 (one). The findings affected six of 155 cases with CR and SPGF, three of 366 men with SPGF only, and one (of 323) normozoospermic subfertile man. The subgroup “CR and SPGF” had over 13-fold enrichment of findings compared to controls (3.9% vs. 0.3%; Fisher’s exact test},
note = {DOI: 10.3389/fendo.2024.1312357},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
RASopathies are syndromes caused by congenital defects in the Ras/mitogen-activated protein kinase (MAPK) pathway genes, with a population prevalence of 1 in 1,000. Patients are typically identified in childhood based on diverse characteristic features, including cryptorchidism (CR) in >50% of affected men. As CR predisposes to spermatogenic failure (SPGF; total sperm count per ejaculate 0–39 million), we hypothesized that men seeking infertility management include cases with undiagnosed RASopathies. Likely pathogenic or pathogenic (LP/P) variants in 22 RASopathy-linked genes were screened in 521 idiopathic SPGF patients (including 155 CR cases) and 323 normozoospermic controls using exome sequencing. All 844 men were recruited to the ESTonian ANDrology (ESTAND) cohort and underwent identical andrological phenotyping. RASopathy-specific variant interpretation guidelines were used for pathogenicity assessment. LP/P variants were identified in PTPN11 (two), SOS1 (three), SOS2 (one), LZTR1 (one), SPRED1 (one), NF1 (one), and MAP2K1 (one). The findings affected six of 155 cases with CR and SPGF, three of 366 men with SPGF only, and one (of 323) normozoospermic subfertile man. The subgroup “CR and SPGF” had over 13-fold enrichment of findings compared to controls (3.9% vs. 0.3%; Fisher’s exact test |
Stefanija, Ana Pop; Buelens, Bart; Goesaert, Elfi; Lenaerts, Tom; Pierson, Jean René; den Bussche, Jan Van Toward a Solid Acceptance of the Decentralized Web of Personal Data: Societal and Technological Convergence Journal Article In: Communications of the ACM, vol. 67, no. 1, pp. 43-46, 2024, (DOI: 10.1145/3624555). @article{info:hdl:2013/367025,
title = {Toward a Solid Acceptance of the Decentralized Web of Personal Data: Societal and Technological Convergence},
author = {Ana Pop Stefanija and Bart Buelens and Elfi Goesaert and Tom Lenaerts and Jean René Pierson and Jan Van den Bussche},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/367025/4/3624555.pdf},
year = {2024},
date = {2024-01-01},
journal = {Communications of the ACM},
volume = {67},
number = {1},
pages = {43-46},
abstract = {Giving individuals more control of their personal data.},
note = {DOI: 10.1145/3624555},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Giving individuals more control of their personal data. |
Lenaerts, Tom; Saponara, Marco; Pacheco, Jorge J. M.; Santos, Francisco C. Evolution of a theory of mind Journal Article In: iScience, vol. 27, no. 2, 2024, (DOI: 10.1016/j.isci.2024.108862). @article{info:hdl:2013/372022,
title = {Evolution of a theory of mind},
author = {Tom Lenaerts and Marco Saponara and Jorge J. M. Pacheco and Francisco C. Santos},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/372022/1/doi_355666.pdf},
year = {2024},
date = {2024-01-01},
journal = {iScience},
volume = {27},
number = {2},
abstract = {Even though the Theory of Mind in upper primates has been under investigation for decades, how it may evolve remains an open problem. We propose here an evolutionary game theoretical model where a finite population of individuals may use reasoning strategies to infer a response to the anticipated behavior of others within the context of a sequential dilemma, i.e., the Centipede Game. We show that strategies with bounded reasoning evolve and flourish under natural selection, provided they are allowed to make reasoning mistakes and a temptation for higher future gains is in place. We further show that non-deterministic reasoning co-evolves with an optimism bias that may lead to the selection of new equilibria, closely associated with average behavior observed in experimental data. This work reveals both a novel perspective on the evolution of bounded rationality and a co-evolutionary link between the evolution of Theory of Mind and the emergence of misbeliefs.},
note = {DOI: 10.1016/j.isci.2024.108862},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Even though the Theory of Mind in upper primates has been under investigation for decades, how it may evolve remains an open problem. We propose here an evolutionary game theoretical model where a finite population of individuals may use reasoning strategies to infer a response to the anticipated behavior of others within the context of a sequential dilemma, i.e., the Centipede Game. We show that strategies with bounded reasoning evolve and flourish under natural selection, provided they are allowed to make reasoning mistakes and a temptation for higher future gains is in place. We further show that non-deterministic reasoning co-evolves with an optimism bias that may lead to the selection of new equilibria, closely associated with average behavior observed in experimental data. This work reveals both a novel perspective on the evolution of bounded rationality and a co-evolutionary link between the evolution of Theory of Mind and the emergence of misbeliefs. |
Terrucha, Ines; Domingos, Elias Fernández; Santos, Francisco C.; Simoens, Pieter; Lenaerts, Tom The art of compensation: How hybrid teams solve collective-risk dilemmas Journal Article In: PloS one, vol. 19, no. 2 February, 2024, (DOI: 10.1371/journal.pone.0297213). @article{info:hdl:2013/371876,
title = {The art of compensation: How hybrid teams solve collective-risk dilemmas},
author = {Ines Terrucha and Elias Fernández Domingos and Francisco C. Santos and Pieter Simoens and Tom Lenaerts},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/371876/1/doi_355520.pdf},
year = {2024},
date = {2024-01-01},
journal = {PloS one},
volume = {19},
number = {2 February},
abstract = {It is widely known how the human ability to cooperate has influenced the thriving of our species. However, as we move towards a hybrid human-machine future, it is still unclear how the introduction of artificial agents in our social interactions affect this cooperative capacity. In a one-shot collective risk dilemma, where enough members of a group must cooperate in order to avoid a collective disaster, we study the evolutionary dynamics of cooperation in a hybrid population. In our model, we consider a hybrid population composed of both adaptive and fixed behavior agents. The latter serve as proxies for the machine-like behavior of artificially intelligent agents who implement stochastic strategies previously learned offline. We observe that the adaptive individuals adjust their behavior in function of the presence of artificial agents in their groups to compensate their cooperative (or lack of thereof) efforts. We also find that risk plays a determinant role when assessing whether or not we should form hybrid teams to tackle a collective risk dilemma. When the risk of collective disaster is high, cooperation in the adaptive population falls dramatically in the presence of cooperative artificial agents. A story of compensation, rather than cooperation, where adaptive agents have to secure group success when the artificial agents are not cooperative enough, but will rather not cooperate if the others do so. On the contrary, when risk of collective disaster is low, success is highly improved while cooperation levels within the adaptive population remain the same. Artificial agents can improve the collective success of hybrid teams. However, their application requires a true risk assessment of the situation in order to actually benefit the adaptive population (i.e. the humans) in the long-term.},
note = {DOI: 10.1371/journal.pone.0297213},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
It is widely known how the human ability to cooperate has influenced the thriving of our species. However, as we move towards a hybrid human-machine future, it is still unclear how the introduction of artificial agents in our social interactions affect this cooperative capacity. In a one-shot collective risk dilemma, where enough members of a group must cooperate in order to avoid a collective disaster, we study the evolutionary dynamics of cooperation in a hybrid population. In our model, we consider a hybrid population composed of both adaptive and fixed behavior agents. The latter serve as proxies for the machine-like behavior of artificially intelligent agents who implement stochastic strategies previously learned offline. We observe that the adaptive individuals adjust their behavior in function of the presence of artificial agents in their groups to compensate their cooperative (or lack of thereof) efforts. We also find that risk plays a determinant role when assessing whether or not we should form hybrid teams to tackle a collective risk dilemma. When the risk of collective disaster is high, cooperation in the adaptive population falls dramatically in the presence of cooperative artificial agents. A story of compensation, rather than cooperation, where adaptive agents have to secure group success when the artificial agents are not cooperative enough, but will rather not cooperate if the others do so. On the contrary, when risk of collective disaster is low, success is highly improved while cooperation levels within the adaptive population remain the same. Artificial agents can improve the collective success of hybrid teams. However, their application requires a true risk assessment of the situation in order to actually benefit the adaptive population (i.e. the humans) in the long-term. |
Gravel, Barbara; Renaux, Alexandre; Papadimitriou, Sofia; Smits, Guillaume; Nowe, Ann; Lenaerts, Tom Prioritization of oligogenic variant combinations in whole exomes Journal Article In: Bioinformatics, vol. 40, no. 4, 2024, (DOI: 10.1093/bioinformatics/btae184). @article{info:hdl:2013/374647,
title = {Prioritization of oligogenic variant combinations in whole exomes},
author = {Barbara Gravel and Alexandre Renaux and Sofia Papadimitriou and Guillaume Smits and Ann Nowe and Tom Lenaerts},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/374647/1/doi_358291.pdf},
year = {2024},
date = {2024-01-01},
journal = {Bioinformatics},
volume = {40},
number = {4},
abstract = {Motivation: Whole exome sequencing (WES) has emerged as a powerful tool for genetic research, enabling the collection of a tremendous amount of data about human genetic variation. However, properly identifying which variants are causative of a genetic disease remains an important challenge, often due to the number of variants that need to be screened. Expanding the screening to combinations of variants in two or more genes, as would be required under the oligogenic inheritance model, simply blows this problem out of proportion. Results: We present here the High-throughput oligogenic prioritizer (Hop), a novel prioritization method that uses direct oligogenic information at the variant, gene and gene pair level to detect digenic variant combinations in WES data. This method leverages information from a knowledge graph, together with specialized pathogenicity predictions in order to effectively rank variant combinations based on how likely they are to explain the patient’s phenotype. The performance of Hop is evaluated in cross-validation on 36 120 synthetic exomes for training and 14 280 additional synthetic exomes for independent testing. Whereas the known pathogenic variant combinations are found in the top 20 in approximately 60% of the cross-validation exomes, 71% are found in the same ranking range when considering the independent set. These results provide a significant improvement over alternative approaches that depend simply on a monogenic assessment of pathogenicity, including early attempts for digenic ranking using monogenic pathogenicity scores.},
note = {DOI: 10.1093/bioinformatics/btae184},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Motivation: Whole exome sequencing (WES) has emerged as a powerful tool for genetic research, enabling the collection of a tremendous amount of data about human genetic variation. However, properly identifying which variants are causative of a genetic disease remains an important challenge, often due to the number of variants that need to be screened. Expanding the screening to combinations of variants in two or more genes, as would be required under the oligogenic inheritance model, simply blows this problem out of proportion. Results: We present here the High-throughput oligogenic prioritizer (Hop), a novel prioritization method that uses direct oligogenic information at the variant, gene and gene pair level to detect digenic variant combinations in WES data. This method leverages information from a knowledge graph, together with specialized pathogenicity predictions in order to effectively rank variant combinations based on how likely they are to explain the patient’s phenotype. The performance of Hop is evaluated in cross-validation on 36 120 synthetic exomes for training and 14 280 additional synthetic exomes for independent testing. Whereas the known pathogenic variant combinations are found in the top 20 in approximately 60% of the cross-validation exomes, 71% are found in the same ranking range when considering the independent set. These results provide a significant improvement over alternative approaches that depend simply on a monogenic assessment of pathogenicity, including early attempts for digenic ranking using monogenic pathogenicity scores. |
Lillepea, Kristiina; Juchnewitsch, Anna Grete; Kasak, Laura; Valkna, Anu; Dutta, Avirup; Pomm, Kristjan; Poolamets, Olev; Nagirnaja, Liina; Tamp, Erik; Mahyari, Eisa; Vihljajev, Vladimir; Tjagur, Stanislav; Papadimitriou, Sofia; Riera-Escamilla, Antoni; Versbraegen, Nassim; Farnetani, Ginevra; Castillo-Madeen, Helen; Sütt, Mailis; Kübarsepp, Viljo; Tennisberg, Sven; Korrovits, Paul; Krausz, Csilla; Aston, Kenneth Ivan; Lenaerts, Tom; Conrad, Donald D. F.; Punab, Margus; Laan, Maris Toward clinical exomes in diagnostics and management of male infertility Journal Article In: American journal of human genetics, vol. 111, no. 5, pp. 877-895, 2024, (DOI: 10.1016/j.ajhg.2024.03.013). @article{info:hdl:2013/374767,
title = {Toward clinical exomes in diagnostics and management of male infertility},
author = {Kristiina Lillepea and Anna Grete Juchnewitsch and Laura Kasak and Anu Valkna and Avirup Dutta and Kristjan Pomm and Olev Poolamets and Liina Nagirnaja and Erik Tamp and Eisa Mahyari and Vladimir Vihljajev and Stanislav Tjagur and Sofia Papadimitriou and Antoni Riera-Escamilla and Nassim Versbraegen and Ginevra Farnetani and Helen Castillo-Madeen and Mailis Sütt and Viljo Kübarsepp and Sven Tennisberg and Paul Korrovits and Csilla Krausz and Kenneth Ivan Aston and Tom Lenaerts and Donald D. F. Conrad and Margus Punab and Maris Laan},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/374767/3/Lillepeaetal.pdf},
year = {2024},
date = {2024-01-01},
journal = {American journal of human genetics},
volume = {111},
number = {5},
pages = {877-895},
abstract = {Infertility, affecting ∼10% of men, is predominantly caused by primary spermatogenic failure (SPGF). We screened likely pathogenic and pathogenic (LP/P) variants in 638 candidate genes for male infertility in 521 individuals presenting idiopathic SPGF and 323 normozoospermic men in the ESTAND cohort. Molecular diagnosis was reached for 64 men with SPGF (12%), with findings in 39 genes (6%). The yield did not differ significantly between the subgroups with azoospermia (20/185, 11%), oligozoospermia (18/181, 10%), and primary cryptorchidism with SPGF (26/155, 17%). Notably, 19 of 64 LP/P variants (30%) identified in 28 subjects represented recurrent findings in this study and/or with other male infertility cohorts. NR5A1 was the most frequently affected gene, with seven LP/P variants in six SPGF-affected men and two normozoospermic men. The link to SPGF was validated for recently proposed candidate genes ACTRT1, ASZ1, GLUD2, GREB1L, LEO1, RBM5, ROS1, and TGIF2LY. Heterozygous truncating variants in BNC1, reported in female infertility, emerged as plausible causes of severe oligozoospermia. Data suggested that several infertile men may present congenital conditions with less pronounced or pleiotropic phenotypes affecting the development and function of the reproductive system. Genes regulating the hypothalamic-pituitary-gonadal axis were affected in >30% of subjects with LP/P variants. Six individuals had more than one LP/P variant, including five with two findings from the gene panel. A 4-fold increased prevalence of cancer was observed in men with genetic infertility compared to the general male population (8% vs. 2%; p = 4.4 ?x 10−3). Expanding genetic testing in andrology will contribute to the multidisciplinary management of SPGF.},
note = {DOI: 10.1016/j.ajhg.2024.03.013},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Infertility, affecting ∼10% of men, is predominantly caused by primary spermatogenic failure (SPGF). We screened likely pathogenic and pathogenic (LP/P) variants in 638 candidate genes for male infertility in 521 individuals presenting idiopathic SPGF and 323 normozoospermic men in the ESTAND cohort. Molecular diagnosis was reached for 64 men with SPGF (12%), with findings in 39 genes (6%). The yield did not differ significantly between the subgroups with azoospermia (20/185, 11%), oligozoospermia (18/181, 10%), and primary cryptorchidism with SPGF (26/155, 17%). Notably, 19 of 64 LP/P variants (30%) identified in 28 subjects represented recurrent findings in this study and/or with other male infertility cohorts. NR5A1 was the most frequently affected gene, with seven LP/P variants in six SPGF-affected men and two normozoospermic men. The link to SPGF was validated for recently proposed candidate genes ACTRT1, ASZ1, GLUD2, GREB1L, LEO1, RBM5, ROS1, and TGIF2LY. Heterozygous truncating variants in BNC1, reported in female infertility, emerged as plausible causes of severe oligozoospermia. Data suggested that several infertile men may present congenital conditions with less pronounced or pleiotropic phenotypes affecting the development and function of the reproductive system. Genes regulating the hypothalamic-pituitary-gonadal axis were affected in >30% of subjects with LP/P variants. Six individuals had more than one LP/P variant, including five with two findings from the gene panel. A 4-fold increased prevalence of cancer was observed in men with genetic infertility compared to the general male population (8% vs. 2%; p = 4.4 ?x 10−3). Expanding genetic testing in andrology will contribute to the multidisciplinary management of SPGF. |
Nachtegael, Charlotte; Stefani, Jacopo De; Cnudde, Anthony; Lenaerts, Tom DUVEL: an active-learning annotated biomedical corpus for the recognition of oligogenic combinations Journal Article In: Database, vol. 2024, no. 2024, 2024, (DOI: 10.1093/database/baae039). @article{info:hdl:2013/374632,
title = {DUVEL: an active-learning annotated biomedical corpus for the recognition of oligogenic combinations},
author = {Charlotte Nachtegael and Jacopo De Stefani and Anthony Cnudde and Tom Lenaerts},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/374632/1/doi_358276.pdf},
year = {2024},
date = {2024-01-01},
journal = {Database},
volume = {2024},
number = {2024},
abstract = {Abstract While biomedical relation extraction (bioRE) datasets have been instrumental in the development of methods to support biocuration of single variants from texts, no datasets are currently available for the extraction of digenic or even oligogenic variant relations, despite the reports in literature that epistatic effects between combinations of variants in different loci (or genes) are important to understand disease etiologies. This work presents the creation of a unique dataset of oligogenic variant combinations, geared to train tools to help in the curation of scientific literature. To overcome the hurdles associated with the number of unlabelled instances and the cost of expertise, active learning (AL) was used to optimize the annotation, thus getting assistance in finding the most informative subset of samples to label. By pre-annotating 85 full-text articles containing the relevant relations from the Oligogenic Diseases Database (OLIDA) with PubTator, text fragments featuring potential digenic variant combinations, i.e. gene–variant–gene–variant, were extracted. The resulting fragments of texts were annotated with ALAMBIC, an AL-based annotation platform. The resulting dataset, called DUVEL, is used to fine-tune four state-of-the-art biomedical language models: BiomedBERT, BiomedBERT-large, BioLinkBERT and BioM-BERT. More than 500 000 text fragments were considered for annotation, finally resulting in a dataset with 8442 fragments, 794 of them being positive instances, covering 95% of the original annotated articles. When applied to gene–variant pair detection, BiomedBERT-large achieves the highest F1 score (0.84) after fine-tuning, demonstrating significant improvement compared to the non-fine-tuned model, underlining the relevance of the DUVEL dataset. This study shows how AL may play an important role in the creation of bioRE dataset relevant for biomedical curation applications. DUVEL provides a unique biomedical corpus focusing on 4-ary relations between two genes and two variants. It is made freely available for research on GitHub and Hugging Face. Database URL: https://huggingface.co/datasets/cnachteg/duvel or https://doi.org/10.57967/hf/1571},
note = {DOI: 10.1093/database/baae039},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Abstract While biomedical relation extraction (bioRE) datasets have been instrumental in the development of methods to support biocuration of single variants from texts, no datasets are currently available for the extraction of digenic or even oligogenic variant relations, despite the reports in literature that epistatic effects between combinations of variants in different loci (or genes) are important to understand disease etiologies. This work presents the creation of a unique dataset of oligogenic variant combinations, geared to train tools to help in the curation of scientific literature. To overcome the hurdles associated with the number of unlabelled instances and the cost of expertise, active learning (AL) was used to optimize the annotation, thus getting assistance in finding the most informative subset of samples to label. By pre-annotating 85 full-text articles containing the relevant relations from the Oligogenic Diseases Database (OLIDA) with PubTator, text fragments featuring potential digenic variant combinations, i.e. gene–variant–gene–variant, were extracted. The resulting fragments of texts were annotated with ALAMBIC, an AL-based annotation platform. The resulting dataset, called DUVEL, is used to fine-tune four state-of-the-art biomedical language models: BiomedBERT, BiomedBERT-large, BioLinkBERT and BioM-BERT. More than 500 000 text fragments were considered for annotation, finally resulting in a dataset with 8442 fragments, 794 of them being positive instances, covering 95% of the original annotated articles. When applied to gene–variant pair detection, BiomedBERT-large achieves the highest F1 score (0.84) after fine-tuning, demonstrating significant improvement compared to the non-fine-tuned model, underlining the relevance of the DUVEL dataset. This study shows how AL may play an important role in the creation of bioRE dataset relevant for biomedical curation applications. DUVEL provides a unique biomedical corpus focusing on 4-ary relations between two genes and two variants. It is made freely available for research on GitHub and Hugging Face. Database URL: https://huggingface.co/datasets/cnachteg/duvel or https://doi.org/10.57967/hf/1571 |
Kirchsteiger, Georg; Lenaerts, Tom; Suchon, Remi Voluntary versus mandatory information disclosure in the sequential prisoner’s dilemma Journal Article In: Economic theory, 2024, (DOI: 10.1007/s00199-024-01563-y). @article{info:hdl:2013/373750,
title = {Voluntary versus mandatory information disclosure in the sequential prisoner’s dilemma},
author = {Georg Kirchsteiger and Tom Lenaerts and Remi Suchon},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/373750/3/s00199-024-01563-y.pdf},
year = {2024},
date = {2024-01-01},
journal = {Economic theory},
abstract = {In sequential social dilemmas with stranger matching, initiating cooperation is inherently risky for the first mover. The disclosure of the second mover’s past actions may be necessary to instigate cooperation. We experimentally compare the effect of mandatory and voluntary disclosure with non-disclosure in a sequential prisoner’s dilemma situation. Our results confirm the positive effects of disclosure on cooperation. We also find that voluntary disclosure is as effective as mandatory disclosure, which runs counter to the results of existing literature on this topic. With voluntary disclosure, second movers who have a good track record chose to disclose, suggesting that they anticipate non-disclosure would signal non-cooperativeness. First movers interpret non-disclosure correctly as a signal of non-cooperativeness. Therefore, they cooperate less than half as often when the second mover decides not to disclose.},
note = {DOI: 10.1007/s00199-024-01563-y},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
In sequential social dilemmas with stranger matching, initiating cooperation is inherently risky for the first mover. The disclosure of the second mover’s past actions may be necessary to instigate cooperation. We experimentally compare the effect of mandatory and voluntary disclosure with non-disclosure in a sequential prisoner’s dilemma situation. Our results confirm the positive effects of disclosure on cooperation. We also find that voluntary disclosure is as effective as mandatory disclosure, which runs counter to the results of existing literature on this topic. With voluntary disclosure, second movers who have a good track record chose to disclose, suggesting that they anticipate non-disclosure would signal non-cooperativeness. First movers interpret non-disclosure correctly as a signal of non-cooperativeness. Therefore, they cooperate less than half as often when the second mover decides not to disclose. |
Terrucha, Ines; Domingos, Elias Fernández; Simoens, Pieter; Lenaerts, Tom Committing to the wrong artificial delegate in a collective-risk dilemma is better than directly committing mistakes Journal Article In: Scientific reports, vol. 14, no. 1, 2024, (DOI: 10.1038/s41598-024-61153-9). @article{info:hdl:2013/374814,
title = {Committing to the wrong artificial delegate in a collective-risk dilemma is better than directly committing mistakes},
author = {Ines Terrucha and Elias Fernández Domingos and Pieter Simoens and Tom Lenaerts},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/374814/1/doi_358458.pdf},
year = {2024},
date = {2024-01-01},
journal = {Scientific reports},
volume = {14},
number = {1},
abstract = {While autonomous artificial agents are assumed to perfectly execute the strategies they are programmed with, humans who design them may make mistakes. These mistakes may lead to a misalignment between the humans’ intended goals and their agents’ observed behavior, a problem of value alignment. Such an alignment problem may have particularly strong consequences when these autonomous systems are used in social contexts that involve some form of collective risk. By means of an evolutionary game theoretical model, we investigate whether errors in the configuration of artificial agents change the outcome of a collective-risk dilemma, in comparison to a scenario with no delegation. Delegation is here distinguished from no-delegation simply by the moment at which a mistake occurs: either when programming/choosing the agent (in case of delegation) or when executing the actions at each round of the game (in case of no-delegation). We find that, while errors decrease success rate, it is better to delegate and commit to a somewhat flawed strategy, perfectly executed by an autonomous agent, than to commit execution errors directly. Our model also shows that in the long-term, delegation strategies should be favored over no-delegation, if given the choice.},
note = {DOI: 10.1038/s41598-024-61153-9},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
While autonomous artificial agents are assumed to perfectly execute the strategies they are programmed with, humans who design them may make mistakes. These mistakes may lead to a misalignment between the humans’ intended goals and their agents’ observed behavior, a problem of value alignment. Such an alignment problem may have particularly strong consequences when these autonomous systems are used in social contexts that involve some form of collective risk. By means of an evolutionary game theoretical model, we investigate whether errors in the configuration of artificial agents change the outcome of a collective-risk dilemma, in comparison to a scenario with no delegation. Delegation is here distinguished from no-delegation simply by the moment at which a mistake occurs: either when programming/choosing the agent (in case of delegation) or when executing the actions at each round of the game (in case of no-delegation). We find that, while errors decrease success rate, it is better to delegate and commit to a somewhat flawed strategy, perfectly executed by an autonomous agent, than to commit execution errors directly. Our model also shows that in the long-term, delegation strategies should be favored over no-delegation, if given the choice. |
Rivière, Quentin; Raskin, Virginie; Melo, Romário; Boutet, Stéphanie; Corso, Massimiliano; Defrance, Matthieu; Webb, Alex A. R.; Verbruggen, Nathalie; Anoman, Djoro Armand Effects of light regimes on circadian gene co‐expression networks in Arabidopsis thaliana Journal Article In: Plant Direct, vol. 8, no. 8, 2024, (DOI: 10.1002/pld3.70001). @article{info:hdl:2013/384388,
title = {Effects of light regimes on circadian gene co‐expression networks in Arabidopsis thaliana},
author = {Quentin Rivière and Virginie Raskin and Romário Melo and Stéphanie Boutet and Massimiliano Corso and Matthieu Defrance and Alex A. R. Webb and Nathalie Verbruggen and Djoro Armand Anoman},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/384388/3/Plant-Direct-2024.pdf},
year = {2024},
date = {2024-01-01},
journal = {Plant Direct},
volume = {8},
number = {8},
abstract = {Abstract Light/dark (LD) cycles are responsible for oscillations in gene expression, which modulate several aspects of plant physiology. Those oscillations can persist under constant conditions due to regulation by the circadian oscillator. The response of the transcriptome to light regimes is dynamic and allows plants to adapt rapidly to changing environmental conditions. We compared the transcriptome of Arabidopsis under LD and constant light (LL) for 3 days and identified different gene co‐expression networks in the two light regimes. Our studies yielded unforeseen insights into circadian regulation. Intuitively, we anticipated that gene clusters regulated by the circadian oscillator would display oscillations under LD cycles. However, we found transcripts encoding components of the flavonoid metabolism pathway that were rhythmic in LL but not in LD. We also discovered that the expressions of many stress‐related genes were significantly increased during the dark period in LD relative to the subjective night in LL, whereas the expression of these genes in the light period was similar. The nocturnal pattern of these stress‐related gene expressions suggested a form of “skotoprotection.” The transcriptomics data were made available in a web application named Cyclath , which we believe will be a useful tool to contribute to a better understanding of the impact of light regimes on plants.},
note = {DOI: 10.1002/pld3.70001},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
Abstract Light/dark (LD) cycles are responsible for oscillations in gene expression, which modulate several aspects of plant physiology. Those oscillations can persist under constant conditions due to regulation by the circadian oscillator. The response of the transcriptome to light regimes is dynamic and allows plants to adapt rapidly to changing environmental conditions. We compared the transcriptome of Arabidopsis under LD and constant light (LL) for 3 days and identified different gene co‐expression networks in the two light regimes. Our studies yielded unforeseen insights into circadian regulation. Intuitively, we anticipated that gene clusters regulated by the circadian oscillator would display oscillations under LD cycles. However, we found transcripts encoding components of the flavonoid metabolism pathway that were rhythmic in LL but not in LD. We also discovered that the expressions of many stress‐related genes were significantly increased during the dark period in LD relative to the subjective night in LL, whereas the expression of these genes in the light period was similar. The nocturnal pattern of these stress‐related gene expressions suggested a form of “skotoprotection.” The transcriptomics data were made available in a web application named Cyclath , which we believe will be a useful tool to contribute to a better understanding of the impact of light regimes on plants. |
Jansen, Maarten Information criteria for structured parameter selection in high dimensional tree and graph models Journal Article In: Digital signal processing, vol. 148, 2024, (Language of publication: fr). @article{info:hdl:2013/372845,
title = {Information criteria for structured parameter selection in high dimensional tree and graph models},
author = {Maarten Jansen},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/372845/3/jansen24structuredpreprint.pdf},
year = {2024},
date = {2024-01-01},
journal = {Digital signal processing},
volume = {148},
note = {Language of publication: fr},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
|
Bhattacharya, Shreya; Lefèvre, Laure; Chatzistergos, T; Hayakawa, Hisashi; Jansen, Maarten RudolfWolf to AlfredWolfer: The Transfer of the Reference Observer in the International Sunspot Number Series (1876–1893) Journal Article In: Solar physics, vol. 299, 2024, (Language of publication: fr). @article{info:hdl:2013/372844,
title = {RudolfWolf to AlfredWolfer: The Transfer of the Reference Observer in the International Sunspot Number Series (1876–1893)},
author = {Shreya Bhattacharya and Laure Lefèvre and T Chatzistergos and Hisashi Hayakawa and Maarten Jansen},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/372844/3/PreprintBhattacharya202404.pdf},
year = {2024},
date = {2024-01-01},
journal = {Solar physics},
volume = {299},
note = {Language of publication: fr},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
|
Colot, Martin; Simar, Cédric; Petieau, Mathieu; Alvarez, Ana Maria Cebolla; Chéron, Guy; Bontempi, Gianluca EMG subspace alignment and visualization for cross-subject hand gesture classification Miscellaneous 2024, (Conference: ECML-PKDD 2023 Worshop - Adapting to change : Reliable Learning Across Domains (2023-09-18: Turin)). @misc{info:hdl:2013/373864,
title = {EMG subspace alignment and visualization for cross-subject hand gesture classification},
author = {Martin Colot and Cédric Simar and Mathieu Petieau and Ana Maria Cebolla Alvarez and Guy Chéron and Gianluca Bontempi},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/373864/3/ECML_PKDD_2023__Martin_Colot_Main_and_Appendix.pdf},
year = {2024},
date = {2024-01-01},
abstract = {Electromyograms (EMG)-based hand gesture recognition systems are a promising technology for human/machine interfaces. However, one of their main limitations is the long calibration time that is typically required to handle new users. The paper discusses and analyses the challenge of cross-subject generalization thanks to an original dataset containing the EMG signals of 14 human subjects during hand gestures. The experimental results show that, though an accurate generalization based on pooling multiple subjects is hardly achievable, it is possible to improve the cross-subject estimation by identifying a robust low-dimensional subspace for multiple subjects and aligning it to a target subject. A visualization of the subspace enables us to provide insights for the improvement of cross-subject generalization with EMG signals.},
note = {Conference: ECML-PKDD 2023 Worshop - Adapting to change : Reliable Learning Across Domains (2023-09-18: Turin)},
keywords = {},
pubstate = {published},
tppubtype = {misc}
}
Electromyograms (EMG)-based hand gesture recognition systems are a promising technology for human/machine interfaces. However, one of their main limitations is the long calibration time that is typically required to handle new users. The paper discusses and analyses the challenge of cross-subject generalization thanks to an original dataset containing the EMG signals of 14 human subjects during hand gestures. The experimental results show that, though an accurate generalization based on pooling multiple subjects is hardly achievable, it is possible to improve the cross-subject estimation by identifying a robust low-dimensional subspace for multiple subjects and aligning it to a target subject. A visualization of the subspace enables us to provide insights for the improvement of cross-subject generalization with EMG signals. |
Lebichot, Bertrand; Siblini, Wissam; Paldino, Gian Marco; Borgne, Yann-Aël Le; Oblé, Frédéric; Bontempi, Gianluca Assessment of catastrophic forgetting in continual credit card fraud detection Journal Article In: Expert systems with applications, vol. 249, 2024, (DOI: 10.1016/j.eswa.2024.123445). @article{info:hdl:2013/370795,
title = {Assessment of catastrophic forgetting in continual credit card fraud detection},
author = {Bertrand Lebichot and Wissam Siblini and Gian Marco Paldino and Yann-Aël Le Borgne and Frédéric Oblé and Gianluca Bontempi},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/370795/3/preproofs-main.pdf},
year = {2024},
date = {2024-01-01},
journal = {Expert systems with applications},
volume = {249},
abstract = {The volume of e-commerce continues to increase year after year. Buying goods on the internet is easy and practical, and took a huge boost during the lockdowns of the Covid crisis. However, this is also an open window for fraudsters and the corresponding financial loss costs billions of dollars. In this paper, we study e-commerce credit card fraud detection, in collaboration with our industrial partner, Worldline. Transactional companies are more and more dependent on machine learning models such as deep learning anomaly detection models, as part of real-world fraud detection systems (FDS). We focus on continual learning to find the best model with respect to two objectives: to maximize the accuracy and to minimize the catastrophic forgetting phenomenon. For the latter, we proposed an evaluation procedure to quantify the forgetting in data streams with delayed feedback: the plasticity/stability visualization matrix. We also investigated six strategies and 13 methods on a real-size case study including five months of e-commerce credit card transactions. Finally, we discuss how the trade-off between plasticity and stability is set, in practice, in the case of FDS.},
note = {DOI: 10.1016/j.eswa.2024.123445},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
The volume of e-commerce continues to increase year after year. Buying goods on the internet is easy and practical, and took a huge boost during the lockdowns of the Covid crisis. However, this is also an open window for fraudsters and the corresponding financial loss costs billions of dollars. In this paper, we study e-commerce credit card fraud detection, in collaboration with our industrial partner, Worldline. Transactional companies are more and more dependent on machine learning models such as deep learning anomaly detection models, as part of real-world fraud detection systems (FDS). We focus on continual learning to find the best model with respect to two objectives: to maximize the accuracy and to minimize the catastrophic forgetting phenomenon. For the latter, we proposed an evaluation procedure to quantify the forgetting in data streams with delayed feedback: the plasticity/stability visualization matrix. We also investigated six strategies and 13 methods on a real-size case study including five months of e-commerce credit card transactions. Finally, we discuss how the trade-off between plasticity and stability is set, in practice, in the case of FDS. |
Paldino, Gian Marco; Lebichot, Bertrand; Borgne, Yann-Aël Le; Siblini, Wissam; Oblé, Frédéric; Boracchi, Giacomo; Bontempi, Gianluca The role of diversity and ensemble learning in credit card fraud detection Journal Article In: Advances in Data Analysis and Classification, vol. 18, no. 1, pp. 193-217, 2024, (DOI: 10.1007/s11634-022-00515-5). @article{info:hdl:2013/372242,
title = {The role of diversity and ensemble learning in credit card fraud detection},
author = {Gian Marco Paldino and Bertrand Lebichot and Yann-Aël Le Borgne and Wissam Siblini and Frédéric Oblé and Giacomo Boracchi and Gianluca Bontempi},
url = {https://dipot.ulb.ac.be/dspace/bitstream/2013/372242/1/elsevier_355886.pdf},
year = {2024},
date = {2024-01-01},
journal = {Advances in Data Analysis and Classification},
volume = {18},
number = {1},
pages = {193-217},
abstract = {The number of daily credit card transactions is inexorably growing: the e-commerce market expansion and the recent constraints for the Covid-19 pandemic have significantly increased the use of electronic payments. The ability to precisely detect fraudulent transactions is increasingly important, and machine learning models are now a key component of the detection process. Standard machine learning techniques are widely employed, but inadequate for the evolving nature of customers behavior entailing continuous changes in the underlying data distribution. his problem is often tackled by discarding past knowledge, despite its potential relevance in the case of recurrent concepts. Appropriate exploitation of historical knowledge is necessary: we propose a learning strategy that relies on diversity-based ensemble learning and allows to preserve past concepts and reuse them for a faster adaptation to changes. In our experiments, we adopt several state-of-the-art diversity measures and we perform comparisons with various other learning approaches. We assess the effectiveness of our proposed learning strategy on extracts of two real datasets from two European countries, containing more than 30 M and 50 M transactions, provided by our industrial partner, Worldline, a leading company in the field.},
note = {DOI: 10.1007/s11634-022-00515-5},
keywords = {},
pubstate = {published},
tppubtype = {article}
}
The number of daily credit card transactions is inexorably growing: the e-commerce market expansion and the recent constraints for the Covid-19 pandemic have significantly increased the use of electronic payments. The ability to precisely detect fraudulent transactions is increasingly important, and machine learning models are now a key component of the detection process. Standard machine learning techniques are widely employed, but inadequate for the evolving nature of customers behavior entailing continuous changes in the underlying data distribution. his problem is often tackled by discarding past knowledge, despite its potential relevance in the case of recurrent concepts. Appropriate exploitation of historical knowledge is necessary: we propose a learning strategy that relies on diversity-based ensemble learning and allows to preserve past concepts and reuse them for a faster adaptation to changes. In our experiments, we adopt several state-of-the-art diversity measures and we perform comparisons with various other learning approaches. We assess the effectiveness of our proposed learning strategy on extracts of two real datasets from two European countries, containing more than 30 M and 50 M transactions, provided by our industrial partner, Worldline, a leading company in the field. |