Related papers: MAESTRO: Meta-learning Adaptive Estimation of Scal…
Large Language Models (LLMs) show potential as sequential decision-making agents, but their application is often limited due to a reliance on large, computationally expensive models. This creates a need to improve smaller models, yet…
Large language model (LLM) alignment via reinforcement learning from human preferences (RLHF) suffers from unstable policy updates, ambiguous gradient directions, poor interpretability, and high gradient variance in mainstream pairwise…
In this work, we study the problem of finding Pareto optimal policies in multi-agent reinforcement learning problems with cooperative reward structures. We show that any algorithm where each agent only optimizes their reward is subject to…
Supervised fine-tuning (SFT) has emerged as a crucial method for aligning large language models (LLMs) with human-annotated demonstrations. However, SFT, being an off-policy approach similar to behavior cloning, often struggles with…
We consider an agent who is involved in a Markov decision process and receives a vector of outcomes every round. Her objective is to maximize a global concave reward function on the average vectorial outcome. The problem models applications…
We study the theoretical aspects of Reinforced Language Models (RLMs) from a bi-objective optimization perspective. Specifically, we consider the RLMs as a Pareto optimization problem that maximizes the two conflicting objectives, i.e.,…
A key challenge in applying reinforcement learning (RL) to diffusion large language models (dLLMs) lies in the intractability of their likelihood functions, which are essential for the RL objective, necessitating corresponding approximation…
Large language models (LLMs) fine-tuned with alignment techniques, such as reinforcement learning from human feedback, have been instrumental in developing some of the most capable AI systems to date. Despite their success, existing methods…
Alignment of Large Language Models (LLMs) aims to align outputs with human preferences, and personalized alignment further adapts models to individual users. This relies on personalized reward models that capture user-specific preferences…
Group Relative Policy Optimization (GRPO) enhances LLM reasoning but often induces overconfidence, where incorrect responses yield lower perplexity than correct ones, degrading relative calibration as described by the Area Under the Curve…
Group Relative Policy Optimization (GRPO) has emerged as a popular algorithm for reinforcement learning with large language models (LLMs). However, upon analyzing its clipping mechanism, we argue that it is suboptimal in certain scenarios.…
Reinforcement Learning from Human Feedback (RLHF) plays a crucial role in aligning large language models (LLMs) with human values and preferences. However, the quality and stability of the trained reward model largely determine the final…
Large language models (LLMs) used for multiple-choice and pairwise evaluation tasks often exhibit selection bias due to non-semantic factors like option positions and label symbols. Existing inference-time debiasing is costly and may harm…
Multimodal language models now integrate text, audio, and video for unified reasoning. Yet existing RL post-training pipelines treat all input signals as equally relevant, ignoring which modalities each task actually requires. This…
Tables contain rich structured information, yet when stored as images their contents remain "locked" within pixels. Converting table images into LaTeX code enables faithful digitization and reuse, but current multimodal large language…
With the rapid advancement of large language models and vision-language models, employing large models as Web Agents has become essential for automated web interaction. However, training Web Agents with reinforcement learning faces critical…
In this paper, we introduce a Group Relative Policy Optimization (GRPO)-based method for training Speech-Aware Large Language Models (SALLMs) on open-format speech understanding tasks, such as Spoken Question Answering and Automatic Speech…
Prior work in multi-objective reinforcement learning typically uses linear reward scalarization with fixed weights, which provably fails to capture non-convex Pareto fronts and thus yields suboptimal results. This limitation becomes…
Meta-Black-Box Optimization (MetaBBO) is an emerging avenue within Optimization community, where algorithm design policy could be meta-learned by reinforcement learning to enhance optimization performance. So far, the reward functions in…
Reinforcement Learning Fine-Tuning (RLFT) has achieved notable success in tasks with objectively verifiable answers (e.g., code generation, mathematical reasoning), yet struggles with open-ended subjective tasks like role-playing dialogue.…