{"type":"ScholarlyArticle","title":"Quantifying Consistency and Information Loss for Causal Abstraction Learning","authors":[{"name":"Fabio Massimo Zennaro","slug":"fabio-massimo-zennaro","given":"Fabio Massimo","family":"Zennaro","affiliation":"University of Warwick , Coventry , United Kingdom"},{"name":"Paolo Turrini","slug":"paolo-turrini","given":"Paolo","family":"Turrini","affiliation":"University of Warwick , Coventry , United Kingdom"},{"name":"Theodoros Damoulas","slug":"theodoros-damoulas","given":"Theodoros","family":"Damoulas","affiliation":"University of Warwick , Coventry , United Kingdom"}],"abstract":"Structural causal models provide a formalism to express causal relations between variables of interest. Models and variables can represent a system at different levels of abstraction, whereby relations may be coarsened and refined according to the need of a modeller.\r\nHowever, switching between different levels of abstraction requires evaluating a trade-off between the consistency and the information loss among different models.\r\nIn this paper we introduce a family of interventional measures that an agent may use to evaluate such a trade-off. We consider four measures suited for different tasks, analyze their properties, and propose algorithms to evaluate and learn causal abstractions. Finally, we illustrate the flexibility of our setup by empirically showing how different measures and algorithmic choices may lead to different abstractions.","doi":"10.24963/ijcai.2023/638","url":"https://ijcai.org/proceedings/2023/638","pdf":"https://ijcai.org/proceedings/2023/0638.pdf","markdown":"https://ijcai.org/proceedings/2023/638.md","bibtex":"https://ijcai.org/proceedings/2023/bibtex/638","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":"Uncertainty in AI","track":"Main Track","number":638,"pages":{"first":5750,"last":5757},"keywords":["Uncertainty in AI: UAI: Causality, structural causal models and causal inference"],"videos":[],"references":[{"position":1,"raw":"[Beckers and Halpern, 2019] Sander Beckers and Joseph Y Halpern. Abstracting causal models. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 33, pages 2678-2685, 2019.","title":"Abstracting causal models","authors":"Halpern Beckers, Sander Beckers, Joseph Y Halpern","year":"2019","venue":"Proceedings of the AAAI Conference on Artificial Intelligence","doi":null,"url":null,"internal_url":null},{"position":2,"raw":"[Beckers et al., 2020] Sander Beckers, Frederick Eberhardt, and Joseph Y Halpern. Approximate causal abstractions. In Uncertainty in Artificial Intelligence, pages 606-615.","title":"Approximate causal abstractions","authors":"Sander Beckers, Frederick Eberhardt, Joseph Y Halpern","year":null,"venue":"Uncertainty in Artificial Intelligence","doi":null,"url":null,"internal_url":null},{"position":3,"raw":"PMLR, 2020.","title":null,"authors":"Pmlr","year":"2020","venue":null,"doi":null,"url":null,"internal_url":null},{"position":4,"raw":"[Chalupka et al., 2017] Krzysztof Chalupka, Frederick Eber- hardt, and Pietro Perona. Causal feature learning: an overview. Behaviormetrika, 44(1):137-164, 2017.","title":"Causal feature learning: an overview","authors":"Krzysztof Chalupka, Frederick Eberhardt, Pietro Perona","year":"2017","venue":"Behaviormetrika","doi":null,"url":null,"internal_url":null},{"position":5,"raw":"[Chang and Fung, 1990] Kuo-Chu Chang and Robert M. Fung. Refinement and coarsening of bayesian networks. In Conference on Uncertainty in Artificial Intelligence, 1990.","title":"Refinement and coarsening of bayesian networks","authors":"Kuo-Chu Chang, Robert M Fung","year":"1990","venue":"Conference on Uncertainty in Artificial Intelligence","doi":null,"url":null,"internal_url":null},{"position":6,"raw":"[Dietterich, 2000] Thomas G Dietterich. Ensemble methods in machine learning. In International workshop on multi- ple classifier systems, pages 1-15. Springer, 2000.","title":"Ensemble methods in machine learning","authors":"G Thomas, Dietterich","year":"2000","venue":"International workshop on multiple classifier systems","doi":null,"url":null,"internal_url":null},{"position":7,"raw":"[Eberhardt and Lee, 2022] Frederick Eberhardt and Lin Lin Lee. Causal emergence: When distortions in a map ob- scure the territory. Philosophies, 7(2):30, 2022.","title":"Causal emergence: When distortions in a map obscure the territory","authors":"Frederick Eberhardt, Lin Lin, Lee","year":"2022","venue":"Philosophies","doi":null,"url":null,"internal_url":null},{"position":8,"raw":"[Elidan and Gould, 2008] Gal Elidan and Stephen Gould. Learning bounded treewidth bayesian networks. Advances in neural information processing systems, 21, 2008.","title":"Learning bounded treewidth bayesian networks","authors":"Gould Elidan, Gal Elidan, Stephen Gould","year":"2008","venue":"Advances in neural information processing systems","doi":null,"url":null,"internal_url":null},{"position":9,"raw":"[Fritz, 2020] Tobias Fritz. A synthetic approach to markov kernels, conditional independence and theorems on suf- ficient statistics. Advances in Mathematics, 370:107239, 2020.","title":"A synthetic approach to markov kernels, conditional independence and theorems on sufficient statistics","authors":"Tobias Fritz, Fritz","year":"2020","venue":"Advances in Mathematics","doi":null,"url":null,"internal_url":null},{"position":10,"raw":"[Geiger et al., 2021] Atticus Geiger, Hanson Lu, Thomas Icard, and Christopher Potts. Causal abstractions of neu- ral networks. Advances in Neural Information Processing Systems, 34:9574-9586, 2021.","title":"Causal abstractions of neural networks","authors":"Geiger","year":"2021","venue":"Advances in Neural Information Processing Systems","doi":null,"url":null,"internal_url":null},{"position":11,"raw":"[Guyon et al., 2008] Isabelle Guyon, Constantin Aliferis, Greg Cooper, André Elisseeff, Jean-Philippe Pellet, Peter Spirtes, and Alexander Statnikov. Design and analysis of the causation and prediction challenge. In Causation and Prediction Challenge, pages 1-33. PMLR, 2008.","title":"Design and analysis of the causation and prediction challenge","authors":"Isabelle Guyon, Constantin Aliferis, Greg Cooper, André Elisseeff, Jean-Philippe Pellet, Peter Spirtes, Alexander Statnikov","year":"2008","venue":"Causation and Prediction Challenge","doi":null,"url":null,"internal_url":null},{"position":12,"raw":"[Hoel, 2017] Erik P Hoel. When the map is better than the territory. Entropy, 19(5):188, 2017.","title":"When the map is better than the territory","authors":"Erik P Hoel","year":"2017","venue":"Entropy","doi":null,"url":null,"internal_url":null},{"position":13,"raw":"[Kroer and Sandholm, 2018] Christian Kroer and Tuomas Sandholm. A unified framework for extensive-form game abstraction with bounds. In Samy Bengio, Hanna M. Wal- lach, Hugo Larochelle, Kristen Grauman, Nicolò Cesa- Bianchi, and Roman Garnett, editors, Advances in Neu- ral Information Processing Systems 31: Annual Confer- ence on Neural Information Processing Systems 2018, NeurIPS 2018, December 3-8, 2018, Montréal, Canada, pages 613-624, 2018.","title":"Christian Kroer and Tuomas Sandholm. A unified framework for extensive-form game abstraction with bounds","authors":"Sandholm Kroer","year":"2018","venue":"Advances in Neural Information Processing Systems 31: Annual Conference on Neural Information Processing Systems 2018, NeurIPS 2018","doi":null,"url":null,"internal_url":null},{"position":14,"raw":"[Kusner et al., 2017] Matt J Kusner, Joshua Loftus, Chris Russell, and Ricardo Silva. Counterfactual fairness. Ad- vances in neural information processing systems, 30, 2017.","title":"Counterfactual fairness","authors":"Kusner","year":"2017","venue":"Advances in neural information processing systems","doi":null,"url":null,"internal_url":null},{"position":15,"raw":"[Levin, 1992] Simon A Levin. The problem of pattern and scale in ecology: the robert h. macarthur award lecture. Ecology, 73(6):1943-1967, 1992.","title":"The problem of pattern and scale in ecology: the robert h. macarthur award lecture","authors":"Levin Simon","year":"1992","venue":"Ecology","doi":null,"url":null,"internal_url":null},{"position":16,"raw":"[Mitchell, 2021] Melanie Mitchell. Abstraction and analogy- making in artificial intelligence. Annals of the New York Academy of Sciences, 1505(1):79-101, 2021.","title":"Abstraction and analogymaking in artificial intelligence","authors":"Melanie Mitchell","year":"2021","venue":"Annals of the New York Academy of Sciences","doi":null,"url":null,"internal_url":null},{"position":17,"raw":"[Otsuka and Saigo, 2022] Jun Otsuka and Hayato Saigo. On the equivalence of causal models: A category-theoretic ap- proach. arXiv preprint arXiv:2201.06981, 2022.","title":"On the equivalence of causal models: A category-theoretic approach","authors":"Saigo Otsuka, Jun Otsuka, Hayato Saigo","year":"2022","venue":null,"doi":null,"url":null,"internal_url":null},{"position":18,"raw":"[Pearl, 2009] Judea Pearl. Causality. Cambridge University Press, 2009.","title":"Judea Pearl. Causality","authors":"Pearl","year":"2009","venue":null,"doi":null,"url":null,"internal_url":null},{"position":19,"raw":"[Rischel and Weichwald, 2021] Eigil F Rischel and Sebas- tian Weichwald. Compositional abstraction error and a cat- egory of causal models. arXiv preprint arXiv:2103.15758, 2021.","title":"Compositional abstraction error and a category of causal models","authors":"F Eigil, Sebastian Rischel, Weichwald","year":"2021","venue":null,"doi":null,"url":null,"internal_url":null},{"position":20,"raw":"[Rischel, 2020] Eigil Fjeldgren Rischel. The category theory of causal models. Master's thesis, University of Copen- hagen, 2020.","title":"Eigil Fjeldgren Rischel. The category theory of causal models","authors":"Rischel","year":"2020","venue":null,"doi":null,"url":null,"internal_url":null},{"position":21,"raw":"[Rokach, 2019] Lior Rokach. Ensemble learning: pattern classification using ensemble methods. World Scientific, 2019.","title":"Ensemble learning: pattern classification using ensemble methods","authors":"Lior Rokach, Rokach","year":"2019","venue":null,"doi":null,"url":null,"internal_url":null},{"position":22,"raw":"[Rubenstein et al., 2017] Paul K Rubenstein, Sebastian We- ichwald, Stephan Bongers, Joris M Mooij, Dominik Janz- ing, Moritz Grosse-Wentrup, and Bernhard Schölkopf. Causal consistency of structural equation models. In 33rd Conference on Uncertainty in Artificial Intelligence (UAI 2017), pages 808-817.","title":"Causal consistency of structural equation models","authors":"Sebastian Paul K Rubenstein, Stephan Weichwald, Bongers, M Joris, Dominik Mooij, Moritz Janzing, Bernhard Grosse-Wentrup, Schölkopf","year":null,"venue":"33rd Conference on Uncertainty in Artificial Intelligence (UAI 2017)","doi":null,"url":null,"internal_url":null},{"position":23,"raw":"Curran Associates, Inc., 2017.","title":null,"authors":"Curran Associates, Inc","year":"2017","venue":null,"doi":null,"url":null,"internal_url":null},{"position":24,"raw":"[Sandholm, 2015] Tuomas Sandholm. Abstraction for solv- ing large incomplete-information games. In Blai Bonet and Sven Koenig, editors, Proceedings of the Twenty- Ninth AAAI Conference on Artificial Intelligence, January 25-30, 2015, Austin, Texas, USA, pages 4127-4131. AAAI Press, 2015.","title":"Abstraction for solving large incomplete-information games","authors":"Tuomas Sandholm, Sandholm","year":"2015","venue":"Proceedings of the Twenty-Ninth AAAI Conference on Artificial Intelligence","doi":null,"url":null,"internal_url":null},{"position":25,"raw":"[Schmutz et al., 2020] Valentin Schmutz, Wulfram Gerstner, and Tilo Schwalger. Mesoscopic population equations for spiking neural networks with synaptic short-term plastic- ity. The Journal of Mathematical Neuroscience, 10(1):1- 32, 2020.","title":"Mesoscopic population equations for spiking neural networks with synaptic short-term plasticity","authors":"Valentin Schmutz, Wulfram Gerstner, Tilo Schwalger","year":"2020","venue":"The Journal of Mathematical Neuroscience","doi":null,"url":null,"internal_url":null},{"position":26,"raw":"[Spivak, 2014] David I Spivak. Category theory for the sci- ences. MIT Press, 2014.","title":"Category theory for the sciences","authors":"I David, Spivak","year":"2014","venue":null,"doi":null,"url":null,"internal_url":null},{"position":27,"raw":"[Zennaro et al., 2023] Fabio Massimo Zennaro, Máté Drávucz, Geanina Apachitei, W. Dhammika Widanage, and Theodoros Damoulas. Jointly learning consistent causal abstractions over multiple interventional distri- butions. In 2nd Conference on Causal Learning and Reasoning, 2023.","title":"Jointly learning consistent causal abstractions over multiple interventional distributions","authors":"Fabio Massimo Zennaro, Máté Drávucz, Geanina Apachitei, W Dhammika Widanage, Theodoros Damoulas","year":"2023","venue":"2nd Conference on Causal Learning and Reasoning","doi":null,"url":null,"internal_url":null}],"license_note":"Copyright International Joint Conferences on Artificial Intelligence."}