I am currently a PhD candidate at McCombs School of Business, UT Austin advised by Maytal Saar-Tsechansky. I received my bachelor’s degree in Statistics at School of the Gifted Young, University of Science and Technology of China and master’s degree in Statistics at Univerisity of Michigan.

Ex-Intern at Netflix Research (advised by James Mclnerney and Nathan Kallus), HBS (advised by Himabindu Lakkaraju), IBM Research (advised by Wei Sun, Max Biggs, and Markus Ettl), Tencent, and Amazon.

Research

My research interests include the broad area of robust machine learning, reinforcement learning (the specific area of contextual bandit), active learning, AI ethics and causal inference. Many of my research aim to answer these questions:

1. How to develop algorithms that can better utilize human expertise to enhance machine learning systems or aid human decisions?
2. How to assess individual preference or algorithm performance in the presence of missing counterfactuals?
3. How to design better human-AI systems and comprehend their sociotechnical implications?

News

• I will join UT Dallas Naveen Jindal School of Management (with a courtesy appointment in CS) as an assistant professor this fall!
• Our paper Nonparametric Discrete Choice Experiments with Machine Learning Guided Adaptive Design is accepted at RealML @ NeurIPS, 2023!
• Two papers - Contextual Recourse Bandits and Confounding-Robust Policy Improvement with Human-AI Teams are accepted at CIST, 2023 and INFORMS Data Science Workshop, 2023, respectively!
• Our paper On the Impact of Algorithmic Recourse on Social Segregation is accepted at ICML, 2023!
• I will join Netflix as a machine learning research intern this summer!
• Our paper Probabilistic Conformal Prediction Using Conditional Random Samples is accepted at AISTATS, 2023 (Preliminary version accepted as Spotlight presentation at ICML Distribution-Free Uncertainty Quantification, 2022)!
• Two papers - Active Incentive Learning / Robust Human-AI Collaboration with Bandit Feedback (Best Student Paper, 1 out of ~200) are accepted at CIST 2022!
• Our paper Loss Functions for Discrete Contextual Pricing with Observational Data has been selected as a finalist for PhD Incubator Special Recognition Award at INFORMS Advances in Decision Analysis Conference!
• Our paper Loss Functions for Discrete Contextual Pricing with Observational Data is accepted as Spotlight presentation at Revenue Management and Pricing conference, 2022!
• I received UT Austin Graduate School Fellowship for 2022-2023!
• Our paper Counterfactual Self-Training is accepted at AAAI, 2022! (Acceptance Rate: 15%)
• Our paper AE-StyleGAN is accepted at WACV, 2022!
• Our paper P2GAN is accepted at ICCV, 2021!
• Our paper Human-AI Collaboration with Bandit Feedback is accepted at IJCAI, 2021! (Acceptance Rate: 13.9%)

Under Review at Journals

• R Gao, M Saar-Tsechansky, M De-Arteaga, L Han, MK Lee, W Sun, M Lease. Learning Complementary Policies for Human-AI Teams (Under review at Management Science - Reject and Resubmit, previous title: Robust Human-AI Collaboration with Bandit Feedback, Best Student Paper at CIST, 2022)
• M Biggs*, R Gao*, W Sun* Loss Functions for Discrete Contextual Pricing with Observational Data (Under review at Operations Research - Major Revision, spotlight presentation at RMP 2022, special recognition award finalist at ADA 2022).

Selected Publications at ML/AI Conferences

• R Gao, H Lakkaraju. On the Impact of Algorithmic Recourse on Social Segregation (ICML, 2023)
• Z Wang*, R Gao*, M Yin*, M Zhou, D Blei. Probabilistic Conformal Prediction Using Conditional Random Samples (AISTATS 2023, preliminary version accepted as spotlight presentation at ICML DFUQ, 2022)
• R Gao, M Biggs, W Sun, and L Han. Enhancing counterfactual classification performance via self-training (AAAI 2022)
• R Gao, M Saar-Tsechansky, M De-Arteaga, L Han, MK Lee, M Lease. Human-AI Collaboration with Bandit Feedback (IJCAI 2021)
• R Gao, M Saar-Tsechansky. Cost-accuracy aware adaptive labeling for active learning. (AAAI 2020)
• L Han, R Gao, M Kim, X Tao, B Liu, D Metaxas. Robust Conditional GAN from Uncertainty-Aware Pairwise Comparisons. (AAAI 2020)
• L Han, Y Zou, R Gao, L Wang, D Metaxas. Unsupervised domain adaptation via calibrating uncertainties. (CVPR Workshop 2019)
• L Han, MR Min, A Stathopoulos, Y Tian, R Gao, A Kadav, D Metaxas. Dual Projection Generative Adversarial Networks for Conditional Image Generation. (ICCV 2021)

Working Journal Papers

Manuscript will be shared upon request.

• R Gao, M Saar-Tsechansky. Active Incentive Learning (Preliminary version accepted at CIST, 2022). [Abstract]
• R Gao, M Yin. Confounding-Robust Policy Improvement with Human-AI Teams (Preliminary version accepted at INFORMS Data Science Workshop, 2023). [Abstract]
• J Cao*, R Gao*, E Keyvanshokooh*. Contextual Recourse Bandits: Optimizing Decisions through Counterfactual Explanations (Preliminary version accepted at CIST, 2023). [Abstract]
• M Yin, R Gao, W Lin, S Shugan. Nonparametric Discrete Choice Experiments with Machine Learning Guided Adaptive Design (Preliminary version is accepted at RealML @ NeurIPS, 2023). [Abstract]

Patent

• R Gao, W Sun, M Biggs, M Ettl, Y Drissi. Counterfactual Self-Training US Patent App. 17/402,367.
• R Gao, W Sun, M Biggs, Y Drissi, M Ettl. Imputing Counterfactual Data to Facilitate Machine Learning Model Training US Patent App. 17/654,617.

* Equal Contribution

Awards and Fellowships

• Best Student Paper Award at CIST 2022 (1 out of ~200).
• PhD Incubator Special Recognition Award at INFORMS Advances in Decision Analysis Conference.
• INFORMS Data Science Workshop Scholarship 2022, 2023.
• UT Austin Continuing Fellowship 2022-2023 (competitive fellowship, one nomination per department).
• UT Austin Graduate School (OGS) Professional Development Award, Good Systems Student Conference Grant.
• UT Austin Graduate School (OGS) Provost Fellowship.

Professional Service

Program Committee Member and Reviewer for ICLR 2024, WITS 2023, ICIS 2023, NeurIPS 2023, FAccT 2023, AISTATS 2023/2024, AAAI 2023/2024, WITS 2022, CIST 2022, WACV 2022, INFORMS Data Science Workshop, NeurIPS Workshop on Regulatable ML, ICML workshops on Interactive Learning from Human Feedback, Generative AI + Law, Adversarial Machine Learning Frontiers, ML for Data, Human-Machine Collaboration and Teaming..