DISTR combines a task-conditioned diffusion model that generates whole high-return trajectories with behavior-cloning replay, and reports higher average success than several baselines on Continual World.
t-DGR: A trajectory-based deep generative replay method for continual learning in decision making
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Stable Continual Reinforcement Learning via Diffusion-based Trajectory Replay
DISTR combines a task-conditioned diffusion model that generates whole high-return trajectories with behavior-cloning replay, and reports higher average success than several baselines on Continual World.