Qiang Liu

Papers

Total Citations

H-Index

About

Qiang Liu is a prominent researcher whose work spans reinforcement learning, probabilistic inference, and statistical machine learning, with a particular focus on bridging theoretical rigor and practical applicability. His most notable contributions center on off-policy evaluation and estimation methods for reinforcement learning, addressing one of the field's most pressing challenges: how to reliably assess policy performance without direct environment interaction. His 2020 paper on black-box off-policy estimation for infinite-horizon reinforcement learning tackles this problem in high-stakes domains such as healthcare and robotics, where running experiments online is costly or ethically prohibitive. By developing methods that bypass the need for high-fidelity simulators, Liu's research opens pathways for deploying reinforcement learning in real-world settings with greater confidence and safety. With growing citation counts reflecting the community's engagement with his ideas, Liu has established himself as a meaningful voice in advancing the mathematical foundations of sequential decision-making under uncertainty. His work is particularly valuable for students and practitioners seeking principled, model-free approaches to policy optimization and evaluation in complex, long-horizon environments.