Qinghua Liu

Qinghua Liu's work centers on the theory and algorithms of reinforcement learning, especially in settings with partial observability (where an agent cannot fully see the state of its environment) and multi-agent games. Recent work extends this into understanding the capabilities of transformer-based language models, including their memory, planning, and ability to model hidden-state processes like Hidden Markov Models. Some publications also touch on applied machine learning topics such as agricultural image analysis and multimodal sentiment analysis, suggesting occasional collaboration outside the core theoretical focus.

Publication output grew from a low base around 2018 to a peak in 2022, then dipped slightly before rising again in 2025, indicating an active but somewhat variable publication pace over the last decade.

Generated by claude-sonnet-5 from public bibliographic data · Jul 20, 2026