10 papers
MARVL: Multi-Stage Guidance for Robotic Manipulation via Vision-Language Models
Xunlan Zhou, Xuanlin Chen, Shaowei Zhang +4
Designing dense reward functions is pivotal for efficient robotic Reinforcement Learning (RL). However, most dense rewards rely on manual engineering, which fundamentally limits th…
FOUNDER: Grounding Foundation Models in World Models for Open-Ended Embodied Decision Making
Yucen Wang, Rui Yu, Shenghua Wan +2
Foundation Models (FMs) and World Models (WMs) offer complementary strengths in task generalization at different levels. In this work, we propose FOUNDER, a framework that integrat…
Reward Models in Deep Reinforcement Learning: A Survey
Rui Yu, Shenghua Wan, Yucen Wang +4
In reinforcement learning (RL), agents continually interact with the environment and use the feedback to refine their behavior. To guide policy optimization, reward models are intr…
MOS: Model Surgery for Pre-Trained Model-Based Class-Incremental Learning
Hai-Long Sun, Da-Wei Zhou, Hanbin Zhao +3
Class-Incremental Learning (CIL) requires models to continually acquire knowledge of new classes without forgetting old ones. Despite Pre-trained Models (PTMs) have shown excellent…
Weight Scope Alignment: A Frustratingly Easy Method for Model Merging
Yichu Xu, Xin-Chun Li, Le Gan +1
Merging models becomes a fundamental procedure in some applications that consider model efficiency and robustness. The training randomness or Non-I.I.D. data poses a huge challenge…
SeMOPO: Learning High-quality Model and Policy from Low-quality Offline Visual Datasets
Shenghua Wan, Ziyuan Chen, Le Gan +2
Model-based offline reinforcement Learning (RL) is a promising approach that leverages existing data effectively in many real-world applications, especially those involving high-di…
Exploring Dark Knowledge under Various Teacher Capacities and Addressing Capacity Mismatch
Wen-Shu Fan, Xin-Chun Li, De-Chuan Zhan
Knowledge Distillation (KD) could transfer the ``dark knowledge" of a well-performed yet large neural network to a weaker but lightweight one. From the view of output logits and so…
SENSOR: Imitate Third-Person Expert's Behaviors via Active Sensoring
Kaichen Huang, Minghao Shao, Shenghua Wan +4
In many real-world visual Imitation Learning (IL) scenarios, there is a misalignment between the agent's and the expert's perspectives, which might lead to the failure of imitation…
DIDA: Denoised Imitation Learning based on Domain Adaptation
Kaichen Huang, Hai-Hang Sun, Shenghua Wan +4
Imitating skills from low-quality datasets, such as sub-optimal demonstrations and observations with distractors, is common in real-world applications. In this work, we focus on th…
AD3: Implicit Action is the Key for World Models to Distinguish the Diverse Visual Distractors
Yucen Wang, Shenghua Wan, Le Gan +2
Model-based methods have significantly contributed to distinguishing task-irrelevant distractors for visual control. However, prior research has primarily focused on heterogeneous…