I am a fourth year undergraduate at Tokyo Institute of Technology.
My research lies at the intersection of machine learning and causal inference called counterfactual machine learning. I am interested in the counterfactual nature of logged bandit feedback obtained from interactive systems, and ways of using biased real-world datasets to assist better decision making. Most recently, I have been focusing on the intersection of counterfactual machine learning and information retrieval.
- Counterfactual Machine Learning
- Offline evaluation of bandit policies
- Debiasing recommender systems
- Unbiased Learning-to-Rank
- Causal Inference
- Information Retrieval
- Unsupervised Domain Adaptation
- 2020.08: We release Open Bandit Dataset (a large-scale dataset for bandit algorithms) and Open Bandit Pipeline (python package for bandit algorithms and off-policy evaluation).
- 2020.07: My solo paper: “Doubly Robust Estimator for Ranking Metrics with Post-Click Conversions” has been accepted by RecSys2020.
- 2020.06: Our paper: “Unbiased Lift-based Bidding System” has been accepted by AdKDD2020.
- 2020.06: My solo paper: “Unbiased Pairwise Learning from Biased Implicit Feedback” has been accepted by ICTIR2020.
- 2020.06: Our paper: “Counterfactual Cross-Validation: Stable Model Selection Procedure for Causal Inference Models” has been accepted by ICML2020.
- 2020.04: My solo paper: “Asymmetric Tri-training for Debiasing Missing-Not-At-Random Explicit Feedback” has been accepted by SIGIR2020.
- 2020.04: Our paper: “ Dual Learning Algorithm for Delayed Conversions” has been accepted by SIGIR2020.
- 2019.12: Our paper: “Cost-Effective and Stable Policy Optimization Algorithm for Uplift Modeling with Multiple Treatments” has been accepted by SDM2020.
- 2019.10: Our paper “Unbiased Recommender Learning from Missing-Not-At-Random Implicit Feedback” has been accepted by WSDM2020.
- 2018.12: Our paper “Doubly Robust Prediction and Evaluation Methods Improve Uplift Modeling for Observational Data” has been accepted by SDM2019.