This project asks whether a reinforcement learning agent can learn a good patch-selection policy from sparse classification reward, and whether augmenting the policy with a forward-model prediction error signal (either as intrinsic reward or as an observation feature) improves over a vanilla baseline.
Experimental results demonstrate that adaptability is not solely determined by model size, but rather by how effectively parameter plasticity is regulated in dynamic environments.
Xiao-Rong Zeng, Weiqiang Chen, Peng Shi et al.· 0 citations
This work proposes masked boundary modeling, a self-supervised paradigm that dynamically learns sub-pixel boundary representations and subsequently leverages the discovered boundary-bearing tokens as masked targets to facilitate dense visual token learning.
Zelin Fu, Bin Tan, Chang Sun et al.· 3 citations· ⚡2
This paper explores post-training reinforcement learning (RL), specifically GRPO, to directly align autoregressive perception models with their evaluation metrics, and designs an RL framework that addresses perception-specific challenges: reward design for set-structured outputs and multi-head sampling control.
Sofian Chaybouti, Yasser Dahou, N. Huynh et al.· 0 citations
This work believes Patch Policy provides a pipeline for the robotics community to readily leverage continuing progress in visual representation learning, without sacrificing the training efficiency or inference speed required for high-frequency, reactive control.
Gaoyue Zhou, Zichen Jeff Cui, Ada Langford et al.· arXiv.org· 0 citations
A coherent map of the rapidly expanding landscape of visual RL is provided to provide researchers and practitioners with a coherent map of the rapidly expanding landscape of visual RL and to highlight promising directions for future inquiry.
