How I Study AI - Learn AI Papers & Lectures the Easy Way

MAXS: Meta-Adaptive Exploration with LLM Agents

Intermediate

Jian Zhang, Zhiyuan Wang et al.Jan 14arXiv

MAXS is a new way for AI agents to think a few steps ahead while using tools like search and code, so they make smarter choices.

#LLM agents#tool-augmented reasoning#lookahead

JudgeRLVR: Judge First, Generate Second for Efficient Reasoning

Intermediate

Jiangshan Duo, Hanyu Li et al.Jan 13arXiv

JudgeRLVR teaches a model to be a strict judge of answers before it learns to generate them, which trims bad ideas early.

#RLVR#judge-then-generate#discriminative supervision

GDPO: Group reward-Decoupled Normalization Policy Optimization for Multi-reward RL Optimization

Intermediate

Shih-Yang Liu, Xin Dong et al.Jan 8arXiv

When a model learns from many rewards at once, a popular method called GRPO can accidentally squash different reward mixes into the same learning signal, which confuses training.

#GDPO#GRPO#multi-reward reinforcement learning

Papers3

MAXS: Meta-Adaptive Exploration with LLM Agents

JudgeRLVR: Judge First, Generate Second for Efficient Reasoning

GDPO: Group reward-Decoupled Normalization Policy Optimization for Multi-reward RL Optimization