{"type":"ScholarlyArticle","title":"Towards Theoretically Grounded Evolutionary Learning","authors":[{"name":"Chao Qian","slug":"chao-qian","given":"Chao","family":"Qian","affiliation":"State Key Laboratory for Novel Software Technology , Nanjing University , Nanjing 210023 , China"}],"abstract":"Machine learning tasks are often formulated as complex optimization problems, where the objective function can be non-differentiable, non-continuous, non-unique, inaccurate, dynamic, and have many local optima, making conventional optimization algorithms fail. Evolutionary Algorithms (EAs), inspired by Darwin's theory of evolution, are general-purpose randomized heuristic optimization algorithms, mimicking variational reproduction and natural selection. EAs have yielded encouraging outcomes for solving complex optimization problems (e.g., neural architecture search) in machine learning. However, due to the heuristic nature of EAs, most outcomes to date have been empirical and lack theoretical support, encumbering their acceptance to the general machine learning community. In this paper, I will review the progress towards theoretically grounded evolutionary learning, from the aspects of analysis methodology, theoretical perspectives and learning algorithms. Due to space limit, I will include a few representative examples and highlight our contributions. I will also discuss some future challenges.","doi":"10.24963/ijcai.2022/819","url":"https://ijcai.org/proceedings/2022/819","pdf":"https://ijcai.org/proceedings/2022/0819.pdf","markdown":"https://ijcai.org/proceedings/2022/819.md","bibtex":"https://ijcai.org/proceedings/2022/bibtex/819","volume":{"slug":"2022","title":"Proceedings of the Thirty-First International Joint Conference on Artificial Intelligence","short_title":"IJCAI 2022","url":"https://ijcai.org/proceedings/2022/","year":2022,"location":"Vienna","dates":"23-29 July 2022","editors":"Luc De Raedt","isbn":"978-1-956792-00-3","publisher":"International Joint Conferences on Artificial Intelligence Organization","doi_prefix":"10.24963/ijcai.2022"},"section":"Early Career","track":"Early Career","number":819,"pages":{"first":5826,"last":5830},"keywords":["EC: Evolutionary Computation","EC: Machine Learning"],"videos":[{"label":"Video #1","url":"https://slideslive.com/38987223","length_sec":90,"provider":"slideslive"},{"label":"Video #2","url":"https://slideslive.com/38985366","length_sec":899,"provider":"slideslive"}],"references":[{"position":1,"raw":"[Bäck, 1996] T. Bäck. Evolutionary Algorithms in Theory and Practice: Evolution Strategies, Evolutionary Programming, Ge- netic Algorithms. Oxford University Press, 1996.","title":"Evolutionary Algorithms in Theory and Practice: Evolution Strategies, Evolutionary Programming, Genetic Algorithms","authors":"T Bäck","year":"1996","venue":null,"doi":null,"url":null,"internal_url":null},{"position":2,"raw":"[Bian and Qian, 2022] C. Bian and C. Qian. Better running time of the non-dominated sorting genetic algorithm ii (NSGA-II) by using stochastic tournament selection. In PPSN, 2022.","title":"Better running time of the non-dominated sorting genetic algorithm ii (NSGA-II) by using stochastic tournament selection","authors":"C Bian, C Qian","year":"2022","venue":"PPSN","doi":null,"url":null,"internal_url":null},{"position":3,"raw":"[Bian et al., 2018] C. Bian, C. Qian, and K. Tang. A general ap- proach to running time analysis of multi-objective evolutionary algorithms. In IJCAI, 2018.","title":"A general approach to running time analysis of multi-objective evolutionary algorithms","authors":"C Bian, C Qian, K Tang","year":"2018","venue":"IJCAI","doi":null,"url":null,"internal_url":"/proceedings/2018/195"},{"position":4,"raw":"[Bian et al., 2021] C. Bian, C. Qian, F. Neumann, and Y. Yu. Fast Pareto optimization for subset selection with dynamic cost con- straints. In IJCAI, 2021.","title":"Fast Pareto optimization for subset selection with dynamic cost constraints","authors":"C Bian, C Qian, F Neumann, Y Yu","year":"2021","venue":"IJCAI","doi":null,"url":null,"internal_url":"/proceedings/2021/302"},{"position":5,"raw":"[Borodin et al., 2017] A. Borodin, A. Jain, H. C. Lee, and Y. Ye. Max-sum diversification, monotone submodular functions, and dynamic updates. ACM TALG, 13(3):1-25, 2017.","title":"Max-sum diversification, monotone submodular functions, and dynamic updates","authors":"A Borodin, A Jain, H C Lee, Y Ye","year":"2017","venue":"ACM TALG","doi":null,"url":null,"internal_url":null},{"position":6,"raw":"[Corus et al., 2017] D. Corus, D.-C. Dang, A. Eremeev, and P. K. Lehre. Level-based analysis of genetic algorithms and other search processes. IEEE TEvC, 22(5):707-719, 2017.","title":"Level-based analysis of genetic algorithms and other search processes","authors":"D Corus, D.-C Dang, A Eremeev, P K Lehre","year":"2017","venue":"IEEE TEvC","doi":null,"url":null,"internal_url":null},{"position":7,"raw":"[Crawford, 2021] V. G. Crawford. Faster guarantees of evolution- ary algorithms for maximization of monotone submodular func- tions. In IJCAI, 2021.","title":"Faster guarantees of evolutionary algorithms for maximization of monotone submodular functions","authors":"V G Crawford","year":"2021","venue":"IJCAI","doi":null,"url":null,"internal_url":"/proceedings/2021/229"},{"position":8,"raw":"[Deb et al., 2002] K. Deb, A. Pratap, S. Agarwal, and T. Meyari- van. A fast and elitist multiobjective genetic algorithm: NSGA- II. IEEE TEvC, 6(2):182-197, 2002.","title":"A fast and elitist multiobjective genetic algorithm: NSGA-II","authors":"K Deb, A Pratap, S Agarwal, T Meyarivan","year":"2002","venue":"IEEE TEvC","doi":null,"url":null,"internal_url":null},{"position":9,"raw":"[Doerr et al., 2012] B. Doerr, D. Johannsen, and C. Winzen. Mul- tiplicative drift analysis. Algorithmica, 64(4):673-697, 2012.","title":"Multiplicative drift analysis","authors":"B Doerr, D Johannsen, C Winzen","year":"2012","venue":"Algorithmica","doi":null,"url":null,"internal_url":null},{"position":10,"raw":"[Domingos, 2012] P. Domingos. A few useful things to know about machine learning. CACM, 55(10):78-87, 2012.","title":"A few useful things to know about machine learning","authors":"P Domingos","year":"2012","venue":"CACM","doi":null,"url":null,"internal_url":null},{"position":11,"raw":"[Friedrich and Neumann, 2015] T. Friedrich and F. Neumann. Maximizing submodular functions under matroid constraints by evolutionary algorithms. ECJ, 23(4):543-558, 2015.","title":"Maximizing submodular functions under matroid constraints by evolutionary algorithms","authors":"T Friedrich, F Neumann","year":"2015","venue":"ECJ","doi":null,"url":null,"internal_url":null},{"position":12,"raw":"[Harshaw et al., 2019] C. Harshaw, M. Feldman, J. Ward, and A. Karbasi. Submodular maximization beyond non-negativity: Guarantees, fast algorithms, and applications. In ICML, 2019.","title":"Submodular maximization beyond non-negativity: Guarantees, fast algorithms, and applications","authors":"C Harshaw, M Feldman, J Ward, A Karbasi","year":"2019","venue":null,"doi":null,"url":null,"internal_url":null},{"position":13,"raw":"[He and Yao, 2001] J. He and X. Yao. Drift analysis and average time complexity of evolutionary algorithms. AIJ, 127(1):57-85, 2001.","title":"Drift analysis and average time complexity of evolutionary algorithms","authors":"J He, X Yao","year":"2001","venue":"AIJ","doi":null,"url":null,"internal_url":null},{"position":14,"raw":"[Horel and Singer, 2016] T. Horel and Y. Singer. Maximization of approximately submodular functions. In NIPS, 2016.","title":"Maximization of approximately submodular functions","authors":"T Horel, Y Singer","year":"2016","venue":"NIPS","doi":null,"url":null,"internal_url":null},{"position":15,"raw":"[Krause et al., 2008] A. Krause, A. Singh, and C. Guestrin. Near- optimal sensor placements in Gaussian processes: Theory, effi- cient algorithms and empirical studies. JMLR, 9:235-284, 2008.","title":"Nearoptimal sensor placements in Gaussian processes: Theory, efficient algorithms and empirical studies","authors":"A Krause, A Singh, C Guestrin","year":"2008","venue":"JMLR","doi":null,"url":null,"internal_url":null},{"position":16,"raw":"[Laumanns et al., 2004] M. Laumanns, L. Thiele, and E. Zitzler. Running time analysis of multiobjective evolutionary algorithms on pseudo-Boolean functions. IEEE TEvC, 8(2):170-182, 2004.","title":"Running time analysis of multiobjective evolutionary algorithms on pseudo-Boolean functions","authors":"M Laumanns, L Thiele, E Zitzler","year":"2004","venue":"IEEE TEvC","doi":null,"url":null,"internal_url":null},{"position":17,"raw":"[Nemhauser and Wolsey, 1978] G. L. Nemhauser and L. A. Wolsey. Best algorithms for approximating the maximum of a submodular set function. MOR, 3(3):177-188, 1978.","title":"Best algorithms for approximating the maximum of a submodular set function","authors":"G L Nemhauser, L A Wolsey","year":"1978","venue":"MOR","doi":null,"url":null,"internal_url":null},{"position":18,"raw":"[Neumann and Wegener, 2006] F. Neumann and I. Wegener. Min- imum spanning trees made easier via multi-objective optimiza- tion. Natural Computing, 5(3):305-319, 2006.","title":"Minimum spanning trees made easier via multi-objective optimization","authors":"F Neumann, I Wegener","year":"2006","venue":"Natural Computing","doi":null,"url":null,"internal_url":null},{"position":19,"raw":"[Neumann and Witt, 2015] F. Neumann and C. Witt. On the run- time of randomized local search and simple evolutionary algo- rithms for dynamic makespan scheduling. In IJCAI, 2015.","title":"On the runtime of randomized local search and simple evolutionary algorithms for dynamic makespan scheduling","authors":"F Neumann, C Witt","year":"2015","venue":"IJCAI","doi":null,"url":null,"internal_url":"/proceedings/2015/526"},{"position":20,"raw":"[Oliveto and Witt, 2011] P. S. Oliveto and C. Witt. Simplified drift analysis for proving lower bounds in evolutionary computation. Algorithmica, 59(3):369-386, 2011.","title":"Simplified drift analysis for proving lower bounds in evolutionary computation","authors":"P S Oliveto, C Witt","year":"2011","venue":"Algorithmica","doi":null,"url":null,"internal_url":null},{"position":21,"raw":"[Qian et al., 2013] C. Qian, Y. Yu, and Z.-H. Zhou. An analysis on recombination in multi-objective evolutionary optimization. AIJ, 204:99-119, 2013.","title":"An analysis on recombination in multi-objective evolutionary optimization","authors":"C Qian, Y Yu, Z.-H Zhou","year":"2013","venue":"AIJ","doi":null,"url":null,"internal_url":null},{"position":22,"raw":"[Qian et al., 2015a] C. Qian, Y. Yu, and Z.-H. Zhou. On constrained Boolean Pareto optimization. In IJCAI, 2015.","title":"On constrained Boolean Pareto optimization","authors":"Qian","year":"2015","venue":"IJCAI","doi":null,"url":null,"internal_url":"/proceedings/2015/61"},{"position":23,"raw":"[Qian et al., 2015b] C. Qian, Y. Yu, and Z.-H. Zhou. Pareto ensem- ble pruning. In AAAI, 2015.","title":"Pareto ensemble pruning","authors":"Qian","year":"2015","venue":"AAAI","doi":null,"url":null,"internal_url":null},{"position":24,"raw":"[Qian et al., 2015c] C. Qian, Y. Yu, and Z.-H. Zhou. Subset selec- tion by Pareto optimization. In NIPS, 2015.","title":"Subset selection by Pareto optimization","authors":"Qian","year":"2015","venue":"NIPS","doi":null,"url":null,"internal_url":null},{"position":25,"raw":"[Qian et al., 2016] C. Qian, J.-C. Shi, Y. Yu, K. Tang, and Z.-H. Zhou. Parallel Pareto optimization for subset selection. In IJCAI, 2016.","title":"Parallel Pareto optimization for subset selection","authors":"C Qian, J.-C Shi, Y Yu, K Tang, Z.-H Zhou","year":"2016","venue":"IJCAI","doi":null,"url":null,"internal_url":"/proceedings/2016/277"},{"position":26,"raw":"[Qian et al., 2017] C. Qian, J.-C. Shi, Y. Yu, K. Tang, and Z.-H. Zhou. Subset selection under noise. In NIPS, 2017.","title":"Subset selection under noise","authors":"C Qian, J.-C Shi, Y Yu, K Tang, Z.-H Zhou","year":"2017","venue":"NIPS","doi":null,"url":null,"internal_url":null},{"position":27,"raw":"[Qian et al., 2018a] C. Qian, Y. Yu, K. Tang, Y. Jin, X. Yao, and Z.-H. Zhou. On the effectiveness of sampling for evolutionary optimization in noisy environments. ECJ, 26(2):237-267, 2018.","title":"On the effectiveness of sampling for evolutionary optimization in noisy environments","authors":"Qian","year":"2018","venue":"ECJ","doi":null,"url":null,"internal_url":null},{"position":28,"raw":"[Qian et al., 2018b] C. Qian, Y. Yu, and Z.-H. Zhou. Analyzing evolutionary optimization in noisy environments. ECJ, 26(1):1- 41, 2018.","title":"Analyzing evolutionary optimization in noisy environments","authors":"Qian","year":"2018","venue":"ECJ","doi":null,"url":null,"internal_url":null},{"position":29,"raw":"[Qian et al., 2019] C. Qian, Y. Yu, K. Tang, X. Yao, and Z.-H. Zhou. Maximizing submodular or monotone approximately sub- modular functions by multi-objective evolutionary algorithms. AIJ, 275:279-294, 2019.","title":"Maximizing submodular or monotone approximately submodular functions by multi-objective evolutionary algorithms","authors":"C Qian, Y Yu, K Tang, X Yao, Z.-H Zhou","year":"2019","venue":"AIJ","doi":null,"url":null,"internal_url":null},{"position":30,"raw":"[Qian et al., 2021] C. Qian, C. Bian, Y. Yu, K. Tang, and X. Yao. Analysis of noisy evolutionary optimization when sampling fails. Algorithmica, 83(4):940-975, 2021.","title":"Analysis of noisy evolutionary optimization when sampling fails","authors":"C Qian, C Bian, Y Yu, K Tang, X Yao","year":"2021","venue":"Algorithmica","doi":null,"url":null,"internal_url":null},{"position":31,"raw":"[Qian et al., 2022] C. Qian, D.-X. Liu, and Z.-H. Zhou. Result di- versification by multi-objective evolutionary algorithms with the- oretical guarantees. AIJ, 309:103737, 2022.","title":"Result diversification by multi-objective evolutionary algorithms with theoretical guarantees","authors":"C Qian, D.-X Liu, Z.-H Zhou","year":"2022","venue":"AIJ","doi":null,"url":null,"internal_url":null},{"position":32,"raw":"[Qian, 2020] C. Qian. Distributed Pareto optimization for large- scale noisy subset selection. IEEE TEvC, 24(4):694-707, 2020.","title":"Distributed Pareto optimization for largescale noisy subset selection","authors":"C Qian","year":"2020","venue":"IEEE TEvC","doi":null,"url":null,"internal_url":null},{"position":33,"raw":"[Real et al., 2017] E. Real, S. Moore, A. Selle, S. Saxena, Y. L. Suematsu, J. Tan, Q. Le, and A. Kurakin. Large-scale evolution of image classifiers. In ICML, 2017.","title":"Large-scale evolution of image classifiers","authors":"E Real, S Moore, A Selle, S Saxena, Y L Suematsu, J Tan, Q Le, A Kurakin","year":"2017","venue":"ICML","doi":null,"url":null,"internal_url":null},{"position":34,"raw":"[Roostapour et al., 2022] V. Roostapour, A. Neumann, F. Neu- mann, and T. Friedrich. Pareto optimization for subset selection with dynamic cost constraints. AIJ, 302:103597, 2022.","title":"Pareto optimization for subset selection with dynamic cost constraints","authors":"V Roostapour, A Neumann, F Neumann, T Friedrich","year":"2022","venue":"AIJ","doi":null,"url":null,"internal_url":null},{"position":35,"raw":"[Shi et al., 2019] F. Shi, M. Schirneck, T. Friedrich, T. Kötzing, and F. Neumann. Reoptimization time analysis of evolutionary algo- rithms on linear functions under dynamic uniform constraints. Algorithmica, 81:828-857, 2019.","title":"Reoptimization time analysis of evolutionary algorithms on linear functions under dynamic uniform constraints","authors":"F Shi, M Schirneck, T Friedrich, T Kötzing, F Neumann","year":"2019","venue":"Algorithmica","doi":null,"url":null,"internal_url":null},{"position":36,"raw":"[Sudholt, 2013] D. Sudholt. A new method for lower bounds on the running time of evolutionary algorithms. IEEE TEvC, 17(3):418- 435, 2013.","title":"A new method for lower bounds on the running time of evolutionary algorithms","authors":"D Sudholt","year":"2013","venue":"IEEE TEvC","doi":null,"url":null,"internal_url":null},{"position":37,"raw":"[Wang et al., 2022] Y. Wang, K. Xue, and C. Qian. Evolutionary diversity optimization with clustering-based selection for rein- forcement learning. In ICLR, 2022.","title":"Evolutionary diversity optimization with clustering-based selection for reinforcement learning","authors":"Y Wang, K Xue, C Qian","year":"2022","venue":"ICLR","doi":null,"url":null,"internal_url":null},{"position":38,"raw":"[Wegener, 2000] I. Wegener. On the expected runtime and the suc- cess probability of evolutionary algorithms. In WG, 2000.","title":"On the expected runtime and the success probability of evolutionary algorithms","authors":"; I Wegener, Wegener","year":"2000","venue":"WG","doi":null,"url":null,"internal_url":null},{"position":39,"raw":"[Yu et al., 2015] Y. Yu, C. Qian, and Z.-H. Zhou. Switch analy- sis for running time analysis of evolutionary algorithms. IEEE TEvC, 19(6):777-792, 2015.","title":"Switch analysis for running time analysis of evolutionary algorithms","authors":"Y Yu, C Qian, Z.-H Zhou","year":"2015","venue":"IEEE TEvC","doi":null,"url":null,"internal_url":null},{"position":40,"raw":"[Zhou et al., 2002] Z.-H. Zhou, J. Wu, and W. Tang. Ensembling neural networks: Many could be better than all. AIJ, 137(1- 2):239-263, 2002.","title":"Ensembling neural networks: Many could be better than all","authors":"Z.-H Zhou, J Wu, W Tang","year":"2002","venue":"AIJ","doi":null,"url":null,"internal_url":null},{"position":41,"raw":"[Zhou et al., 2019] Z.-H. Zhou, Y. Yu, and C. Qian. Evolutionary Learning: Advances in Theories and Algorithms. Springer, 2019.","title":"Evolutionary Learning: Advances in Theories and Algorithms","authors":"Z.-H Zhou, Y Yu, C Qian","year":"2019","venue":null,"doi":null,"url":null,"internal_url":null}],"license_note":"Copyright International Joint Conferences on Artificial Intelligence."}