Jiyang Gao

Papers

Total Citations

214

H-Index

About

Jiyang Gao is a computer vision and machine learning researcher whose work centers on video understanding, action recognition, and predictive modeling in visual systems. He is best known for his contributions to action anticipation — the challenging task of detecting human actions *before* they occur — a capability with significant real-world implications for robotics, autonomous systems, and intelligent surveillance. His most notable work, **"RED: Reinforced Encoder-Decoder Networks for Action Anticipation"** (2017), introduced a reinforcement learning-based framework that moves beyond passive observation to proactively predict future actions from visual sequences. Rather than relying solely on anticipating future frame representations, RED leverages an encoder-decoder architecture trained with reinforcement signals, achieving meaningful advances in predictive accuracy. The paper has accumulated over 191 citations, reflecting its strong influence on the action anticipation subfield. Gao's research sits at an important intersection of deep learning, temporal reasoning, and embodied AI, tackling problems that require machines to reason not just about what is happening, but what *will* happen. His contributions have helped lay foundational groundwork for proactive vision systems, making his work particularly relevant for researchers advancing human-robot interaction and real-time video analysis.