5th Year Master's Thesis Presentation - Allen Zheng

April 24, 2026  4:30PM—6:00PM

Location:
In Person - Newell-Simon 4305

Speaker:
ALLEN ZHENG, Master's Student, Computer Science Department. Carnegie Mellon University

Attention Over The Past for Data Efficiency in RL

We introduce a framework that incorporates attention over past states into reinforcement learning (RL) in two complementary ways. First, we use an attention mechanism over a trajectory buffer of previously visited states to construct a history-aware critic, replacing the standard neural network critic with an estimate computed as an attention-weighted average over stored values. Second, we replace GAE, which is computed through a single trajectory, with an  attention-weighted advantage  that, in addition to stepping forward through time, also steps according to similar states. States more similar to the current one contribute more to the return estimate, providing a smooth, similarity-weighted alternative to the sequential rollout. Together, these two mechanisms reduce variance by pooling signal across similar states rather than relying on a single trajectory.

Thesis Committee
Geoff Gordon (Chair)
Jeff Schneider

Additional Information 
 

For More Information:
amalloy@cs.cmu.edu


Add event to Google
Add event to iCal