11 papers
Stabilizing Off-Policy Training for Long-Horizon LLM Agent via Turn-Level Importance Sampling and Clipping-Triggered Normalization
Chenliang Li, Adel Elmahdy, Alex Boyd +7
Reinforcement learning (RL) algorithms such as PPO and GRPO are widely used to train large language models (LLMs) for multi-turn agentic tasks. However, in off-policy training pipe…
Reinforcing Multi-Turn Reasoning in LLM Agents via Fine-Grained Reward Structure and Credit Assignment
Quan Wei, Siliang Zeng, Chenliang Li +9
Reinforcement Learning (RL) approaches have been wildly used to enhance the reasoning capabilities of Large Language Model (LLM) agents in long-horizon, multi-turn scenarios. Such…
Aligning Frozen LLMs by Reinforcement Learning: An Iterative Reweight-then-Optimize Approach
Xinnan Zhang, Chenliang Li, Siliang Zeng +6
Aligning large language models (LLMs) with human preferences usually requires fine-tuning methods such as RLHF and DPO. These methods directly optimize the model parameters, so the…
Understanding Inverse Reinforcement Learning under Overparameterization: Non-Asymptotic Analysis and Global Optimality
Ruijia Zhang, Siliang Zeng, Chenliang Li +2
The goal of the Inverse reinforcement learning (IRL) task is to identify the underlying reward function and the corresponding optimal policy from a set of expert demonstrations. Wh…
From Demonstrations to Rewards: Alignment Without Explicit Human Preferences
Siliang Zeng, Yao Liu, Huzefa Rangwala +3
One of the challenges of aligning large models with human preferences lies in both the data requirements and the technical complexities of current approaches. Predominant methods,…
Bridging the Training-Inference Gap in LLMs by Leveraging Self-Generated Tokens
Zhepeng Cen, Yao Liu, Siliang Zeng +4
Language models are often trained to maximize the likelihood of the next token given past tokens in the training dataset. However, during inference time, they are utilized differen…