{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/mlic-a-maxsat-based-framework-for-learning","title":"MLIC: A MaxSAT-Based framework for learning interpretable classification rules","arxiv_id":"1812.01843","date":"2018-12-05","proceeding":null,"authors":["Dmitry Malioutov","Kuldeep S. Meel"],"abstract":"The wide adoption of machine learning approaches in the industry, government,\nmedicine and science has renewed the interest in interpretable machine\nlearning: many decisions are too important to be delegated to black-box\ntechniques such as deep neural networks or kernel SVMs. Historically, problems\nof learning interpretable classifiers, including classification rules or\ndecision trees, have been approached by greedy heuristic methods as essentially\nall the exact optimization formulations are NP-hard. Our primary contribution\nis a MaxSAT-based framework, called MLIC, which allows principled search for\ninterpretable classification rules expressible in propositional logic. Our\napproach benefits from the revolutionary advances in the constraint\nsatisfaction community to solve large-scale instances of such problems. In\nexperimental evaluations over a collection of benchmarks arising from practical\nscenarios, we demonstrate its effectiveness: we show that the formulation can\nsolve large classification problems with tens or hundreds of thousands of\nexamples and thousands of features, and to provide a tunable balance of\naccuracy vs. interpretability. Furthermore, we show that in many problems\ninterpretability can be obtained at only a minor cost in accuracy. The primary\nobjective of the paper is to show that recent advances in the MaxSAT literature\nmake it realistic to find optimal (or very high quality near-optimal) solutions\nto large-scale classification problems. The key goal of the paper is to excite\nresearchers in both interpretable classification and in the CP community to\ntake it further and propose richer formulations, and to develop bespoke solvers\nattuned to the problem of interpretable ML.","url_abs":"http://arxiv.org/abs/1812.01843v1","url_pdf":"http://arxiv.org/pdf/1812.01843v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"mlic-a-maxsat-based-framework-for-learning","repo_url":"https://github.com/meelgroup/mlic","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"none","reach":null}],"tasks":[{"task_slug":"machine-learning","task_name":"BIG-bench Machine Learning"},{"task_slug":"classification-1","task_name":"Classification"},{"task_slug":"classification","task_name":"General Classification"},{"task_slug":"interpretable-machine-learning","task_name":"Interpretable Machine Learning"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"syntology_url":null,"atlas_url":"https://app.syntology.ai/?focus=1812.01843","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}