{"type":"ScholarlyArticle","title":"Learning Preference Models with Sparse Interactions of Criteria","authors":[{"name":"Margot Herin","slug":"margot-herin","given":"Margot","family":"Herin","affiliation":"Sorbonne University , CNRS , LIP6 , Paris , France"},{"name":"Patrice Perny","slug":"patrice-perny","given":"Patrice","family":"Perny","affiliation":"Sorbonne University , CNRS , LIP6 , Paris , France"},{"name":"Nataliya Sokolovska","slug":"nataliya-sokolovska","given":"Nataliya","family":"Sokolovska","affiliation":"Sorbonne University , CNRS , LCQB , Paris , France"}],"abstract":"Multicriteria decision making requires defining the result of conflicting and possibly interacting criteria. Allowing criteria interactions in a decision model increases the complexity of the preference learning task due to the combinatorial nature of the possible interactions. In this paper, we propose an approach to learn a decision model in which the interaction pattern is revealed from preference data and kept as simple as possible.  We consider weighted aggregation functions like multilinear utilities or Choquet integrals, admitting representations including non-linear terms measuring the joint benefit or penalty attached to some combinations of criteria. The weighting coefficients known as Möbius masses model positive or negative synergies among criteria. We propose an approach to learn the Möbius masses, based on iterative reweighted least square for sparse recovery, and dualization to improve scalability. This approach is applied to learn sparse representations of the multilinear utility model and conjunctive/disjunctive forms of the discrete Choquet integral from preferences examples, in aggregation problems possibly involving more than 20 criteria.","doi":"10.24963/ijcai.2023/421","url":"https://ijcai.org/proceedings/2023/421","pdf":"https://ijcai.org/proceedings/2023/0421.pdf","markdown":"https://ijcai.org/proceedings/2023/421.md","bibtex":"https://ijcai.org/proceedings/2023/bibtex/421","volume":{"slug":"2023","title":"Proceedings of the Thirty-Second International Joint Conference on Artificial Intelligence","short_title":"IJCAI 2023","url":"https://ijcai.org/proceedings/2023/","year":2023,"location":"Macao, SAR","dates":"19-25 August 2023","editors":"Edith Elkind","isbn":"978-1-956792-03-4","publisher":"International Joint Conferences on Artificial Intelligence Organization","doi_prefix":"10.24963/ijcai.2023"},"section":"Machine Learning","track":"Main Track","number":421,"pages":{"first":3786,"last":3794},"keywords":["Uncertainty in AI: UAI: Decision and utility theory","Knowledge Representation and Reasoning: KRR: Preference modelling and preference-based reasoning","Machine Learning: ML: Learning preferences or rankings"],"videos":[],"references":[{"position":1,"raw":"[Adeyeba et al., 2015] Titilope A. Adeyeba, Derek T. Ander- son, and Timothy C. Havens. Insights and characteriza- tion of l1-norm based sparsity learning of a lexicographi- cally encoded capacity vector for the Choquet integral. In FUZZ-IEEE, pages 1 -7, 2015.","title":"Insights and characterization of l1-norm based sparsity learning of a lexicographically encoded capacity vector for the Choquet integral","authors":"A Titilope, Derek T Adeyeba, Timothy C Anderson, Havens","year":"2015","venue":"FUZZ-IEEE","doi":null,"url":null,"internal_url":null},{"position":2,"raw":"[Ah-Pine et al., 2018] Julien Ah-Pine, Brice Mayag, and An- toine Rolland. Additive bi-capacity by using mathematical programming. In Third International Conference on Algo- rithmic Decision Theory (ADT), pages 15-29, 2018.","title":"Additive bi-capacity by using mathematical programming","authors":"Julien Ah-Pine, Brice Mayag, Antoine Rolland","year":"2018","venue":"Third International Conference on Algorithmic Decision Theory (ADT)","doi":null,"url":null,"internal_url":null},{"position":3,"raw":"[Anderson et al., 2014] Derek T. Anderson, Stanton R. Price, and Timothy C. Havens. Regularization-based learning of the Choquet integral. In FUZZ-IEEE, pages 2519 -2526, 2014.","title":"Regularization-based learning of the Choquet integral","authors":"Derek T Anderson, Stanton R Price, Timothy C Havens","year":"2014","venue":"FUZZ-IEEE","doi":null,"url":null,"internal_url":null},{"position":4,"raw":"[Bach et al., 2012] Francis Bach, Rodolphe Jenatton, Julien Mairal, Guillaume Obozinski, et al. Optimization with sparsity-inducing penalties. Foundations and Trends® in Machine Learning, 4(1):1-106, 2012.","title":"Optimization with sparsity-inducing penalties","authors":"Francis Bach, Rodolphe Jenatton, Julien Mairal, Guillaume Obozinski","year":"2012","venue":"Foundations and Trends® in Machine Learning","doi":null,"url":null,"internal_url":null},{"position":5,"raw":"[Beck, 2015] Amir Beck. On the convergence of alternating minimization for convex programming with applications to iteratively reweighted least squares and decomposition schemes. SIAM Journal on Optimization, 25(1):185-209, 2015.","title":"On the convergence of alternating minimization for convex programming with applications to iteratively reweighted least squares and decomposition schemes","authors":"Amir Beck","year":"2015","venue":"SIAM Journal on Optimization","doi":null,"url":null,"internal_url":null},{"position":6,"raw":"[Beliakov and Wu, 2019] Gleb Beliakov and Jian-Zhang Wu. Learning fuzzy measures from data: simplifica- tions and optimisation strategies. Information Sciences, 494:100-113, 2019.","title":"Learning fuzzy measures from data: simplifications and optimisation strategies","authors":"Wu ; ] Gleb Beliakov, Jian-Zhang Beliakov, Wu","year":"2019","venue":"Information Sciences","doi":null,"url":null,"internal_url":null},{"position":7,"raw":"[Benabbou et al., 2017] Nawal Benabbou, Patrice Perny, and Paolo Viappiani. Incremental elicitation of Choquet ca- pacities for multicriteria choice, ranking and sorting prob- lems. Artificial Intelligence, 246:152-180, 2017.","title":"Incremental elicitation of Choquet capacities for multicriteria choice, ranking and sorting problems","authors":"Nawal Benabbou, Patrice Perny, Paolo Viappiani","year":"2017","venue":"Artificial Intelligence","doi":null,"url":null,"internal_url":null},{"position":8,"raw":"[Bresson et al., 2020] Roman Bresson, Johanne Cohen, Eyke Hüllermeier, Christophe Labreuche, and Michèle Se- bag. Neural representation and learning of hierarchical 2- additive Choquet integrals. In IJCAI, pages 1984-1991, 2020.","title":"Neural representation and learning of hierarchical 2additive Choquet integrals","authors":"Bresson","year":"2020","venue":"IJCAI","doi":null,"url":null,"internal_url":null},{"position":9,"raw":"[Chateauneuf and Jaffray, 1989] Alain Chateauneuf and Jean-Yves Jaffray. Some characterizations of lower probabilities and other monotone capacities through the use of möbius inversion. Mathematical social sciences, 17(3):263-283, 1989.","title":"Some characterizations of lower probabilities and other monotone capacities through the use of möbius inversion","authors":"Jaffray Chateauneuf, Alain Chateauneuf, Jean-Yves Jaffray","year":"1989","venue":"Mathematical social sciences","doi":null,"url":null,"internal_url":null},{"position":10,"raw":"[Daubechies et al., 2010] Ingrid Daubechies, Ronald De- Vore, Massimo Fornasier, and C Sinan Güntürk. Iteratively reweighted least squares minimization for sparse recov- ery. Communications on Pure and Applied Mathematics: A Journal Issued by the Courant Institute of Mathematical Sciences, 63(1):1-38, 2010.","title":"Iteratively reweighted least squares minimization for sparse recovery","authors":"Ingrid Daubechies, Ronald De-Vore, Massimo Fornasier, C Sinan, Güntürk","year":"2010","venue":"Communications on Pure and Applied Mathematics: A Journal Issued by the Courant Institute of Mathematical Sciences","doi":null,"url":null,"internal_url":null},{"position":11,"raw":"[de Oliveira et al., 2022] Henrique Evangelista de Oliveira, Leonardo Tomazeli Duarte, and João Marcos Travassos Romano. Identification of the Choquet integral parame- ters in the interaction index domain by means of sparse modeling. Expert Systems with Applications, 187, 2022.","title":"Identification of the Choquet integral parameters in the interaction index domain by means of sparse modeling","authors":"Henrique Evangelista De Oliveira, Leonardo Tomazeli Duarte, João Marcos, Travassos Romano","year":"2022","venue":"Expert Systems with Applications","doi":null,"url":null,"internal_url":null},{"position":12,"raw":"[Dyer and Sarin, 1979] James S Dyer and Rakesh K Sarin. Measurable multiattribute value functions. Operations re- search, 27(4):810-822, 1979.","title":"Measurable multiattribute value functions","authors":"S James, Dyer, Rakesh K Sarin","year":"1979","venue":"Operations research","doi":null,"url":null,"internal_url":null},{"position":13,"raw":"[Galand and Mayag, 2017] Lucie Galand and Brice Mayag. A heuristic approach to test the compatibility of a prefer- ence information with a choquet integral model. In Algo- rithmic Decision Theory -5th International Conference, ADT, pages 65-80, 2017.","title":"A heuristic approach to test the compatibility of a preference information with a choquet integral model","authors":"Lucie Galand, Brice Mayag","year":"2017","venue":"Algorithmic Decision Theory -5th International Conference","doi":null,"url":null,"internal_url":null},{"position":14,"raw":"[Grabisch and Labreuche, 2010] Michel Grabisch and Christophe Labreuche. A decade of application of the Choquet and Sugeno integrals in multi-criteria decision aid. Annals of Operations Research, 175(1):247-286, 2010.","title":"A decade of application of the Choquet and Sugeno integrals in multi-criteria decision aid","authors":"Michel Grabisch, Christophe Labreuche","year":"2010","venue":"Annals of Operations Research","doi":null,"url":null,"internal_url":null},{"position":15,"raw":"[Grabisch et al., 2003] Michel Grabisch, Christophe Labreuche, and Jean-Claude Vansnick. On the extension of pseudo-boolean functions for the aggregation of interacting criteria. European Journal of Operational Research, 148(1):28-47, 2003.","title":"On the extension of pseudo-boolean functions for the aggregation of interacting criteria","authors":"Michel Grabisch, Christophe Labreuche, Jean-Claude Vansnick","year":"2003","venue":"European Journal of Operational Research","doi":null,"url":null,"internal_url":null},{"position":16,"raw":"[Grabisch et al., 2008] Michel Grabisch, Ivan Kojadinovic, and Patrick Meyer. A review of methods for capacity iden- tification in Choquet integral based multi-attribute utility theory: Applications of the Kappalab R package. Eu- ropean journal of operational research, 186(2):766-785, 2008.","title":"A review of methods for capacity identification in Choquet integral based multi-attribute utility theory: Applications of the Kappalab R package","authors":"Michel Grabisch, Ivan Kojadinovic, Patrick Meyer","year":"2008","venue":"European journal of operational research","doi":null,"url":null,"internal_url":null},{"position":17,"raw":"[Grabisch et al., 2009] Michel Grabisch, Jean-Luc Marichal, Radko Mesiar, and Endre Pap. Aggregation functions, vol- ume 127. Cambridge University Press, 2009.","title":"Aggregation functions","authors":"Michel Grabisch, Jean-Luc Marichal, Radko Mesiar, Endre Pap","year":"2009","venue":null,"doi":null,"url":null,"internal_url":null},{"position":18,"raw":"[Grabisch, 1996] Michel Grabisch. The application of fuzzy integrals in multicriteria decision making. European jour- nal of operational research, 89(3):445-456, 1996.","title":"The application of fuzzy integrals in multicriteria decision making","authors":"Michel Grabisch","year":"1996","venue":"European journal of operational research","doi":null,"url":null,"internal_url":null},{"position":19,"raw":"[Grabisch, 1997] Michel Grabisch. K-order additive discrete fuzzy measures and their representation. Fuzzy sets and systems, 92(2):167-189, 1997.","title":"K-order additive discrete fuzzy measures and their representation","authors":"Michel Grabisch","year":"1997","venue":"Fuzzy sets and systems","doi":null,"url":null,"internal_url":null},{"position":20,"raw":"[Grabisch, 2016] Michel Grabisch. Set functions, games and capacities in decision making. Springer, 2016.","title":"Set functions, games and capacities in decision making","authors":"Michel Grabisch","year":"2016","venue":null,"doi":null,"url":null,"internal_url":null},{"position":21,"raw":"[Grandvalet, 1998] Yves Grandvalet. Least absolute shrink- age is equivalent to quadratic penalization. In Interna- tional Conference on Artificial Neural Networks, pages 201-206. Springer, 1998.","title":"Least absolute shrinkage is equivalent to quadratic penalization","authors":"Yves Grandvalet, Grandvalet","year":"1998","venue":"International Conference on Artificial Neural Networks","doi":null,"url":null,"internal_url":null},{"position":22,"raw":"[Herin et al., 2022a] Margot Herin, Patrice Perny, and Na- taliya Sokolovska. Learning sparse representations of preferences within Choquet expected utility theory. In The 38th Conference on Uncertainty in Artificial Intelligence, 2022.","title":"Learning sparse representations of preferences within Choquet expected utility theory","authors":"Herin","year":"2022","venue":"The 38th Conference on Uncertainty in Artificial Intelligence","doi":null,"url":null,"internal_url":null},{"position":23,"raw":"[Herin et al., 2022b] Margot Herin, Patrice Perny, and Na- taliya Sokolovska. Learning utilities and sparse represen- tations of capacities for multicriteria decision making with the bipolar Choquet integral. In Multidisciplinary Work- shop on Advances in Preference Handling, IJCAI, 2022.","title":"Learning utilities and sparse representations of capacities for multicriteria decision making with the bipolar Choquet integral","authors":"Herin","year":"2022","venue":"Multidisciplinary Workshop on Advances in Preference Handling","doi":null,"url":null,"internal_url":null},{"position":24,"raw":"[Hüllermeier and Tehrani, 2013] Eyke Hüllermeier and Ali Fallah Tehrani. Efficient learning of classifiers based on the 2-additive choquet integral. In Computational Intelligence in Intelligent Data Analysis, pages 17-29, 2013.","title":"Efficient learning of classifiers based on the 2-additive choquet integral","authors":"Tehrani Hüllermeier, Eyke Hüllermeier, Ali Fallah, Tehrani","year":"2013","venue":"Computational Intelligence in Intelligent Data Analysis","doi":null,"url":null,"internal_url":null},{"position":25,"raw":"[Hurwicz, 1951] Leonid Hurwicz. The generalized bayes minimax principle: a criterion for decision making un- der uncertainty. Cowles Comm. Discuss. Paper Stat, 335:1950, 1951.","title":"The generalized bayes minimax principle: a criterion for decision making under uncertainty","authors":"Leonid Hurwicz","year":"1951","venue":"Cowles Comm. Discuss. Paper Stat","doi":null,"url":null,"internal_url":null},{"position":26,"raw":"[Jünger et al., 1993] Michael Jünger, Gerhard Reinelt, and Stefan Thienel. Practical problem solving with cutting plane algorithms in combinatorialoptimization. In Com- binatorial Optimization, 1993.","title":"Practical problem solving with cutting plane algorithms in combinatorialoptimization","authors":"Michael Jünger, Gerhard Reinelt, Stefan Thienel","year":"1993","venue":"Combinatorial Optimization","doi":null,"url":null,"internal_url":null},{"position":27,"raw":"[Kakula et al., 2020] Siva K Kakula, Anthony J Pinar, Tim- othy C Havens, and Derek T Anderson. Choquet integral ridge regression. In 2020 IEEE International Conference on Fuzzy Systems (FUZZ-IEEE), pages 1-8. IEEE, 2020.","title":"Choquet integral ridge regression","authors":"Anthony J Siva K Kakula, Timothy C Pinar, Derek T Havens, Anderson","year":"2020","venue":"IEEE International Conference on Fuzzy Systems (FUZZ-IEEE)","doi":null,"url":null,"internal_url":null},{"position":28,"raw":"[Keeney et al., 1993] Ralph L Keeney, Howard Raiffa, and Richard F Meyer. Decisions with multiple objectives: pref- erences and value trade-offs. Cambridge university press, 1993.","title":"Decisions with multiple objectives: preferences and value trade-offs","authors":"Howard Ralph L Keeney, Richard F Raiffa, Meyer","year":"1993","venue":null,"doi":null,"url":null,"internal_url":null},{"position":29,"raw":"[Owen, 1975] Guillermo Owen. Multilinear extensions and the banzhaf value. Naval research logistics quarterly, 22(4):741-750, 1975.","title":"Multilinear extensions and the banzhaf value","authors":"Guillermo Owen","year":"1975","venue":"Naval research logistics quarterly","doi":null,"url":null,"internal_url":null},{"position":30,"raw":"[Pelegrina et al., 2018] Guilherme D Pelegrina, Michel Gra- bisch, Leonardo T Duarte, and MT Romano. Multilinear model: New issues in capacity identification. In From Multiple Criteria Decision Aid to Preference Learning (DA2PL'2018), 2018.","title":"Multilinear model: New issues in capacity identification","authors":"Michel Guilherme D Pelegrina, Leonardo T Grabisch, Duarte, Romano","year":"2018","venue":"From Multiple Criteria Decision Aid to Preference Learning (DA2PL'2018)","doi":null,"url":null,"internal_url":null},{"position":31,"raw":"[Pelegrina et al., 2020] Guilherme Dean Pelegrina, Leonardo Tomazeli Duarte, Michel Grabisch, and João Marcos Travassos Romano. The multilinear model in multicriteria decision making: The case of 2-additive capacities and contributions to parameter identification. European Journal of Operational Research, 282(3):945- 956, 2020.","title":"The multilinear model in multicriteria decision making: The case of 2-additive capacities and contributions to parameter identification","authors":"Guilherme Dean Pelegrina, Leonardo Tomazeli Duarte, Michel Grabisch, João Marcos, Travassos Romano","year":"2020","venue":"European Journal of Operational Research","doi":null,"url":null,"internal_url":null},{"position":32,"raw":"[Pinar et al., 2017] Anthony J. Pinar, Derek T. Anderson, Timothy C. Havens, Alina Zare, and Titilope Adeyeba. Measures of the Shapley index for learning lower com- plexity fuzzy integrals. Granul. Comput., 2:303 -319, 2017.","title":"Measures of the Shapley index for learning lower complexity fuzzy integrals","authors":"Anthony J Pinar, Derek T Anderson, Timothy C Havens, Alina Zare, Titilope Adeyeba","year":"2017","venue":"Granul. Comput","doi":null,"url":null,"internal_url":null},{"position":33,"raw":"[Schmeidler, 1989] David Schmeidler. Subjective probabil- ity and expected utility without additivity. Econometrica, 57(3):571-587, 1989.","title":"Subjective probability and expected utility without additivity","authors":"David Schmeidler","year":"1989","venue":"Econometrica","doi":null,"url":null,"internal_url":null},{"position":34,"raw":"[Shawe-Taylor et al., 2004] John Shawe-Taylor, Nello Cris- tianini, et al. Kernel methods for pattern analysis. Cam- bridge university press, 2004.","title":"Kernel methods for pattern analysis","authors":"Shawe-Taylor","year":"2004","venue":null,"doi":null,"url":null,"internal_url":null},{"position":35,"raw":"[Tehrani and Hülermeier, 2013] Ali Fallah Tehrani and Eyke Hülermeier. Ordinal Choquistic regression. In EUSFLAT, 2013.","title":"Ordinal Choquistic regression","authors":"Ali Fallah, Tehrani, Eyke Hülermeier","year":"2013","venue":"EUSFLAT","doi":null,"url":null,"internal_url":null},{"position":36,"raw":"[Tehrani et al., 2014] Ali Fallah Tehrani, Marc Strickert, and Eyke Hüllermeier. The choquet kernel for monotone data. In Esann, 2014.","title":"The choquet kernel for monotone data","authors":"Ali Fallah Tehrani, Marc Strickert, Eyke Hüllermeier","year":"2014","venue":"Esann","doi":null,"url":null,"internal_url":null},{"position":37,"raw":"[Tehrani, 2021] Ali Fallah Tehrani. The choquet kernel on the use of regression problem. Information Sciences, 556:256-272, 2021.","title":"The choquet kernel on the use of regression problem","authors":"Ali Fallah, Tehrani","year":"2021","venue":"Information Sciences","doi":null,"url":null,"internal_url":null},{"position":38,"raw":"[Tibshirani, 1996] Robert Tibshirani. Regression shrinkage and selection via the lasso. Journal of the Royal Statistical Society. Series B (methodological), 58(1):267 -88, 1996.","title":"Robert Tibshirani. Regression shrinkage and selection via the lasso","authors":"Tibshirani","year":"1996","venue":"Journal of the Royal Statistical Society. Series B (methodological)","doi":null,"url":null,"internal_url":null},{"position":39,"raw":"[Waegeman et al., 2009] Willem Waegeman, Bernard De Baets, and Luc Boullart. Kernel-based learning methods for preference aggregation. 4OR, 7(2):169-189, 2009.","title":"Kernel-based learning methods for preference aggregation","authors":"Waegeman","year":"2009","venue":null,"doi":null,"url":null,"internal_url":null},{"position":40,"raw":"[Wakker and Deneffe, 1996] Peter Wakker and Daniel Den- effe. Eliciting von neumann-morgenstern utilities when probabilities are distorted or unknown. Management sci- ence, 42(8):1131-1150, 1996.","title":"Peter Wakker and Daniel Deneffe. Eliciting von neumann-morgenstern utilities when probabilities are distorted or unknown","authors":"Deneffe Wakker","year":"1996","venue":"Management science","doi":null,"url":null,"internal_url":null}],"license_note":"Copyright International Joint Conferences on Artificial Intelligence."}