2021 年了，机器/深度学习还有哪些坑比较好挖？ 第1页

提供一个思路：因果推断。

传统的几个深度学习方向（CV/NLP）红利差不多被挖尽了，而推荐得益于各种业界不尽相同的场景尚有挣扎的空间。那最近比较适合刷文的领域，我觉得因果推荐算一个方向，略微小众，却已经引起很多人的兴趣了，而且严格算起来，算是一个小交叉，毕竟很多概念来自计量经济学。因果推断主要解决两个问题，一是因果关系，二是因果关系的影响。跟传统的机器/深度学习解决相关关系不同，因果推断专注于理解变量之间的因果关系。

本人整理的一个论文list，主要聚焦于因果推断里的Uplift。（文末还有一些实用的学习链接）:

Yger, F., Atif, J., & Sugiyama, M. (n.d.). Uplift Modeling from Separate Labels.

Han, M. (n.d.). Uplift-based User Sensitivity Prediction for Coupon Allocation Optimization in E-commerce.

Studies, S. (n.d.). Supporting Information : Meta-learners for Estimating Heterogeneous Treatment Effects using Machine Learning, 1–26. https://doi.org/10.1073/pnas.1804597116

Hirano, K., Imbens, G. W., & Ridder, G. (2003). Efficient estimation of average treatment effects using the estimated propensity score. Econometrica, 71(4), 1161–1189. https://doi.org/10.1111/1468-0262.00442

Rzepakowski, P. (2010). Decision trees for uplift modeling Rzepakowski Marketing campaign example.

Rzepakowski, P. (2010). Decision trees for uplift modeling. 2010 IEEE International Conference on Data Mining, 441–450. https://doi.org/10.1109/ICDM.2010.62

Hill, J. L. (2011). Bayesian nonparametric modeling for causal inference. Journal of Computational and Graphical Statistics BART, 20(1), 217–240. Bayesian Nonparametric Modeling for Causal Inference

Rzepakowski, P., & Jaroszewicz, S. (2012). Uplift modeling in direct marketing. Journal of Telecommunications and Information Technology, 2012(2), 43–50.

Caro, F., & Gallien, J. (2012). Clearance pricing optimization for a fast-fashion retailer. Operations Research, 60(6), 1404–1422. https://doi.org/10.1287/opre.1120.1102

Rzepakowski, P., & Jaroszewicz, S. (2012). Decision trees for uplift modeling with single and multiple treatments. Knowledge and Information Systems, 32(2), 303–327. https://doi.org/10.1007/s10115-011-0434-0

Deng, A., Xu, Y., Kohavi, R., & Walker, T. (2013). Improving the sensitivity of online controlled experiments by utilizing pre-experiment data. WSDM 2013 - Proceedings of the 6th ACM International Conference on Web Search and Data Mining, 123–132. Improving the sensitivity of online controlled experiments by utilizing pre-experiment data

Rzepakowski, P., Jaroszewicz, S., Zhao, Z., Harinen, T., Zhao, Y., Fang, X., … Misra, S. (2015). Recursive partitioning for heterogeneous causal effects. Proceedings of the 17th SIAM International Conference on Data Mining, SDM 2017, 34(27), 1531–1559. https://doi.org/10.1073/pnas.1510489113

Sołtys, M., Jaroszewicz, S., & Rzepakowski, P. (2015). Ensemble methods for uplift modeling. Data Mining and Knowledge Discovery, 29(6), 1531–1559. https://doi.org/10.1007/s10618-014-0383-9

Austin, P. C., & Stuart, E. A. (2015). Moving towards best practice when using inverse probability of treatment weighting (IPTW) using the propensity score to estimate causal treatment effects in observational studies. Statistics in Medicine, 34(28), 3661–3679. https://doi.org/10.1002/sim.6607

Gutierrez, P., & Gérardy, J.-Y. (2016). Causal Inference and Uplift Modeling A review of the literature. JMLR, 67, 1–13.

Athey, S., & Imbens, G. (2016). Recursive partitioning for heterogeneous causal effects. Proceedings of the National Academy of Sciences of the United States of America, 113(27), 7353–7360. Recursive partitioning for heterogeneous causal effects

Zhao, Y., Fang, X., & Simchi-Levi, D. (2017). Uplift modeling with multiple treatments and general response types. Proceedings of the 17th SIAM International Conference on Data Mining, SDM 2017, 588–596. Proceedings of the 2017 SIAM International Conference on Data Mining (SDM)

Li, C., Yan, X., Deng, X., Qi, Y., Chu, W., Song, L., … Xiong, J. (2018). Reinforcement Learning for Uplift Modeling, 1–22. Retrieved from Reinforcement Learning for Uplift Modeling

Diemert, E., Betlei, A., Renaudin, C., & Amini, M.-R. (2018). A Large Scale Benchmark for Uplift Modeling. Proceedings OfAdKDD & TargetAd (ADKDD’18). ACM, 8, 603–621. https://doi.org/10.1145/nnnnnnn.nnnnnnn

Hitsch, Gg. J., & Misra, S. (2018). Heterogeneous Treatment Effects and Optimal Targeting Policy Evaluation. SSRN Electronic Journal, 1–64. https://doi.org/10.2139/ssrn.3111957

Yamane, I., Yger, F., Atif, J., & Sugiyama, M. (2018). Uplift modeling from separate labels. In Advances in Neural Information Processing Systems (Vol. 2018-Decem, pp. 9927–9937). https://doi.org/10.1111/0034-6527.00321

Zhao, K., Hua, J., Yan, L., Zhang, Q., Xu, H., & Yang, C. (2019). A unified framework for marketing budget allocation. In KDD (pp. 1820–1830). https://doi.org/10.1145/3292500.3330700

Zhao, Z., & Harinen, T. (2019). Uplift Modeling for Multiple Treatments with Cost Optimization. Uber DSAA CCF C. Retrieved from Uplift Modeling for Multiple Treatments with Cost Optimization

Sato, M., Sonoda, T., Singh, J., Zhang, Q., Takemori, S., & Ohkuma, T. (2019). Uplif-based evaluation and optimization of recommenders. RecSys 2019, 296–304. https://doi.org/10.1145/3298689.3347018

Nie, X., & Wager, S. (2019). Quasi-Oracle Estimation of Heterogeneous Treatment Effects. R-Learner, (2017), 1–46.

K, R. (2019). Metalearners for estimating heterogeneous treatment effects using machine learning. PNAS Refen_count_uptodate=64. https://doi.org/10.1073/pnas.1804597116

Dasgupta, I., Wang, J., Chiappa, S., Mitrovic, J., Ortega, P., Raposo, D., … Kurth-Nelson, Z. (2019). Causal Reasoning from Meta-reinforcement Learning. Retrieved from http://arxiv.org/abs/1901.08162

Zhu, S., Ng, I., & Chen, Z. (2020). Causal Discovery with Reinforcement Learning. ICLR Huawei Noah’s Ark Lab, 1–17. Retrieved from http://arxiv.org/abs/1906.04477

光喻老哥有一篇很不错的入门文：

知乎的ruocheng老哥也整理了一个很棒的gitlab链接，包含一些论文和实现代码的地址：