ABSTRACT
Methods for offline A/B testing and counterfactual learning are seeing rapid adoption in search and recommender systems, since they allow efficient reuse of existing log data. However, there are fundamental limits to using existing log data alone, since the counterfactual estimators that are commonly used in these methods can have large bias and large variance when the logging policy is very different from the target policy being evaluated. To overcome this limitation, we explore the question of how to design data-gathering policies that most effectively augment an existing dataset of bandit feedback with additional observations for both learning and evaluation. To this effect, this paper introduces Minimum Variance Augmentation Logging (MVAL), a method for constructing logging policies that minimize the variance of the downstream evaluation or learning problem. We explore multiple approaches to computing MVAL policies efficiently, and find that they can be substantially more effective in decreasing the variance of an estimator than naïve approaches.
Supplemental Material
- A. Agarwal, S. Basu, T. Schnabel, and T. Joachims. 2017. Effective Evaluation using Logged Bandit Feedback from Multiple Loggers. In ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD).Google Scholar
- Alekh Agarwal, Daniel Hsu, Satyen Kale, John Langford, Lihong Li, and Robert Schapire. 2014. Taming the Monster: A Fast and Simple Algorithm for Contextual Bandits. In Proceedings of the 31st International Conference on Machine Learning (Proceedings of Machine Learning Research, Vol. 32), Eric P. Xing and Tony Jebara (Eds.). PMLR, Bejing, China, 1638--1646. http://proceedings.mlr.press/v32/agarwalb14.htmlGoogle Scholar
- Shipra Agrawal and Navin Goyal. 2013. Thompson Sampling for Contextual Bandits with Linear Payoffs. In Proceedings of the 30th International Conference on Machine Learning (Proceedings of Machine Learning Research, Vol. 28), Sanjoy Dasgupta and David McAllester (Eds.). PMLR, Atlanta, Georgia, USA, 127--135. http://proceedings.mlr.press/v28/agrawal13.htmlGoogle Scholar
- Alina Beygelzimer and John Langford. 2009. The offset tree for learning with partial labels. In KDD. ACM, 129--138.Google Scholar
- Alberto Bietti, Alekh Agarwal, and John Langford. 2018. A contextual bandit bake-off. arXiv preprint arXiv:1802.04064 (2018).Google Scholar
- Lé on Bottou, Jonas Peters, Joaquin Qui n onero-Candela, Denis X. Charles, D. Max Chickering, Elon Portugaly, Dipankar Ray, Patrice Simard, and Ed Snelson. 2013. Counterfactual Reasoning and Learning Systems: The Example of Computational Advertising. Journal of Machine Learning Research, Vol. 14, 65 (2013), 3207--3260. http://jmlr.org/papers/v14/bottou13a.htmlGoogle ScholarDigital Library
- N. Cesa-Bianchi and G. Lugosi. 2006. Prediction, Learning, and Games. Cambridge University Press.Google Scholar
- Wei Chu, Seung-Taek Park, Todd Beaupre, Nitin Motgi, Amit Phadke, Seinjuti Chakraborty, and Joe Zachariah. 2009. A case study of behavior-driven conjoint analysis on Yahoo! Front Page Today module. In Proceedings of the 15th ACM SIGKDD international conference on Knowledge discovery and data mining. 1097--1104.Google ScholarDigital Library
- JEAN-MARIE CORNUET, JEAN-MICHEL MARIN, Antonietta Mira, and Christian P Robert. 2012. Adaptive multiple importance sampling. Scandinavian Journal of Statistics, Vol. 39, 4 (2012), 798--812.Google ScholarCross Ref
- Miroslav Dud'ik, John Langford, and Lihong Li. 2011. Doubly robust policy evaluation and learning. In Proceedings of the 28th International Conference on International Conference on Machine Learning. 1097--1104.Google Scholar
- Mehrdad Farajtabar, Yinlam Chow, and Mohammad Ghavamzadeh. 2018. More Robust Doubly Robust Off-policy Evaluation. In ICML.Google Scholar
- Zijun Gao, Yanjun Han, Zhimei Ren, and Zhengqing Zhou. 2019. Batched Multi-armed Bandits Problem. In Advances in Neural Information Processing Systems, H. Wallach, H. Larochelle, A. Beygelzimer, F. dtextquotesingle Alché-Buc, E. Fox, and R. Garnett (Eds.), Vol. 32. Curran Associates, Inc. https://proceedings.neurips.cc/paper/2019/file/20f07591c6fcb220ffe637cda29bb3f6-Paper.pdfGoogle Scholar
- Daniel G Horvitz and Donovan J Thompson. 1952. A generalization of sampling without replacement from a finite universe. Journal of the American statistical Association, Vol. 47, 260 (1952), 663--685.Google ScholarCross Ref
- T. Joachims, A. Swaminathan, and M. de Rijke. 2018. Deep Learning with Logged Bandit Feedback. In ICLR.Google Scholar
- Diederik P Kingma and Jimmy Ba. 2014. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980 (2014).Google Scholar
- John Langford and Tong Zhang. 2007. The epoch-greedy algorithm for multi-armed bandits with side information. Advances in neural information processing systems, Vol. 20 (2007), 817--824.Google Scholar
- Lihong Li, Shunbao Chen, Jim Kleban, and Ankur Gupta. 2015. Counterfactual estimation and optimization of click metrics in search engines: A case study. In Proceedings of the 24th International Conference on World Wide Web. 929--934.Google ScholarDigital Library
- Lihong Li, Wei Chu, John Langford, and Robert E Schapire. 2010. A contextual-bandit approach to personalized news article recommendation. In Proceedings of the 19th international conference on World wide web. 661--670.Google ScholarDigital Library
- Ben London and Ted Sandler. 2019. Bayesian Counterfactual Risk Minimization. In ICML. 4125--4133.Google Scholar
- Vinod Nair and Geoffrey E Hinton. 2010. Rectified linear units improve restricted boltzmann machines. In International Conference on Machine Learning.Google ScholarDigital Library
- Art Owen. 2013. Monte Carlo Theory, Methods and Examples. Stanford.Google Scholar
- N. Sachdeva, Yi Su, and T. Joachims. 2020. Off-policy Bandits with Deficient Support. In Proceedings of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD).Google Scholar
- Alex Strehl, John Langford, Lihong Li, and Sham M Kakade. 2011. Learning from Logged Implicit Exploration Data. In NeurIPS.Google Scholar
- Yi Su, Lequn Wang, Michele Santacatterina, and Thorsten Joachims. 2019. CAB: Continuous Adaptive Blending for Policy Evaluation and Learning. In Proceedings of the 36th International Conference on Machine Learning (Proceedings of Machine Learning Research, Vol. 97), Kamalika Chaudhuri and Ruslan Salakhutdinov (Eds.). PMLR, 6005--6014. http://proceedings.mlr.press/v97/su19a.htmlGoogle Scholar
- Adith Swaminathan and Thorsten Joachims. 2015a. Batch Learning from Logged Bandit Feedback through Counterfactual Risk Minimization. Journal of Machine Learning Research, Vol. 16, 52 (2015), 1731--1755. http://jmlr.org/papers/v16/swaminathan15a.htmlGoogle ScholarDigital Library
- A. Swaminathan and T. Joachims. 2015b. The Self-Normalized Estimator for Counterfactual Learning. In NeurIPS.Google Scholar
- A. Swaminathan, A. Krishnamurthy, A. Agarwal, M. Dudik, J. Langford, D. Jose, and I. Zitouni. 2017. Off-policy Evaluation for Slate Recommendation. In Advances in Neural Information Processing Systems (NeurIPS).Google Scholar
- Philip Thomas and Emma Brunskill. 2016. Data-efficient off-policy policy evaluation for reinforcement learning. In International Conference on Machine Learning. PMLR, 2139--2148.Google Scholar
- Hastagiri P. Vanchinathan, Isidor Nikolic, Fabio De Bona, and Andreas Krause. 2014. Explore-Exploit in Top-N Recommender Systems via Gaussian Processes. In Proceedings of the 8th ACM Conference on Recommender Systems (Foster City, Silicon Valley, California, USA) (RecSys '14). Association for Computing Machinery, New York, NY, USA, 225--232. https://doi.org/10.1145/2645710.2645733Google ScholarDigital Library
- Yu-Xiang Wang, Alekh Agarwal, and Miroslav Dudik. 2017. Optimal and Adaptive Off-policy Evaluation in Contextual Bandits. In Proceedings of the 34th International Conference on Machine Learning (Proceedings of Machine Learning Research, Vol. 70). PMLR, International Convention Centre, Sydney, Australia, 3589--3597. http://proceedings.mlr.press/v70/wang17a.htmlGoogle Scholar
- Songbai Yan, Kamalika Chaudhuri, and Tara Javidi. 2018. Active Learning with Logged Data. In Proceedings of the 35th International Conference on Machine Learning (Proceedings of Machine Learning Research, Vol. 80), Jennifer Dy and Andreas Krause (Eds.). PMLR, 5521--5530. https://proceedings.mlr.press/v80/yan18a.htmlGoogle Scholar
- Songbai Yan, Kamalika Chaudhuri, and Tara Javidi. 2019. The Label Complexity of Active Learning from Observational Data. In Advances in Neural Information Processing Systems, H. Wallach, H. Larochelle, A. Beygelzimer, F. d'Alché-Buc, E. Fox, and R. Garnett (Eds.), Vol. 32. Curran Associates, Inc. https://proceedings.neurips.cc/paper/2019/file/1019c8091693ef5c5f55970346633f92-Paper.pdfGoogle Scholar
Index Terms
- Variance-Minimizing Augmentation Logging for Counterfactual Evaluation in Contextual Bandits
Recommendations
Counterfactual Evaluation and Learning for Interactive Systems: Foundations, Implementations, and Recent Advances
KDD '22: Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data MiningCounterfactual estimators enable the use of existing log data to estimate how some new target policy would have performed, if it had been used instead of the policy that logged the data. We say that those estimators work "off-policy", since the policy ...
Contextual Bandits with Cross-Learning
In the classic contextual bandits problem, in each round t, a learner observes some context c, chooses some action i to perform, and receives some reward ri,t(c). We consider the variant of this problem in which in addition to receiving the reward ri,t(c), the learner ...
Off-Policy Evaluation via Adaptive Weighting with Data from Contextual Bandits
KDD '21: Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data MiningIt has become increasingly common for data to be collected adaptively, for example using contextual bandits. Historical data of this type can be used to evaluate other treatment assignment policies to guide future innovation or experiments. However, ...
Comments