Reinforcement Learning
Provable Partially Observable Reinforcement Learning with Privileged Information Yang Cai
Partial observability of the underlying states generally presents significant challenges for reinforcement learning (RL).
Adversarially Trained Weighted Actor-Critic for Safe Offline Reinforcement Learning
Additionally, we offer a practical version of WSAC and compare it with existing state-of-the-art safe offline RL algorithms in several continuous control environments.
5c186016d0844767209dc36e9e61441b-Paper-Conference.pdf
DeMa's focus on sequences diminishes approximately exponentially.
Online Control with Adversarial Disturbance for Continuous-time Linear Systems
A major challenge in robotics is to deploy simulated controllers into real-world.
PrefPaint: Aligning Image Inpainting Diffusion Model with Human Preference
Specifically, instead of directly measuring the divergence with paired images, we train a reward model with the dataset we construct, consisting of nearly 51,000 images annotated with human preferences.
Policy Learning from Tutorial Books via Understanding, Rehearsing and Introspecting Xiong-Hui Chen
However, current research for decision-making, like reinforcement learning (RL), has primarily required numerous real interactions with the target environment to learn a skill, while failing to utilize the existing knowledge already summarized in the text.