Enhancing Multi-Step Reasoning Abilities of Language Models through Direct Q-Function Optimization
- URL: http://arxiv.org/abs/2410.09302v2
- Date: Tue, 11 Feb 2025 01:46:35 GMT
- Title: Enhancing Multi-Step Reasoning Abilities of Language Models through Direct Q-Function Optimization
- Authors: Kaixuan Ji, Guanlin Liu, Ning Dai, Qingping Yang, Renjie Zheng, Zheng Wu, Chen Dun, Quanquan Gu, Lin Yan,
- Abstract summary: Reinforcement Learning (RL) plays a crucial role in aligning large language models with human preferences and improving their ability to perform complex tasks.
We introduce Direct Q-function Optimization (DQO), which formulates the response generation process as a Markov Decision Process (MDP) and utilizes the soft actor-critic (SAC) framework to optimize a Q-function directly parameterized by the language model.
Experimental results on two math problem-solving datasets, GSM8K and MATH, demonstrate that DQO outperforms previous methods, establishing it as a promising offline reinforcement learning approach for aligning language models.
- Score: 49.362750475706235
- License:
- Abstract: Reinforcement Learning (RL) plays a crucial role in aligning large language models (LLMs) with human preferences and improving their ability to perform complex tasks. However, current approaches either require significant computational resources due to the use of multiple models and extensive online sampling for training (e.g., PPO) or are framed as bandit problems (e.g., DPO, DRO), which often struggle with multi-step reasoning tasks, such as math problem solving and complex reasoning that involve long chains of thought. To overcome these limitations, we introduce Direct Q-function Optimization (DQO), which formulates the response generation process as a Markov Decision Process (MDP) and utilizes the soft actor-critic (SAC) framework to optimize a Q-function directly parameterized by the language model. The MDP formulation of DQO offers structural advantages over bandit-based methods, enabling more effective process supervision. Experimental results on two math problem-solving datasets, GSM8K and MATH, demonstrate that DQO outperforms previous methods, establishing it as a promising offline reinforcement learning approach for aligning language models.
Related papers
- Multi-granular Training Strategies for Robust Multi-hop Reasoning Over Noisy and Heterogeneous Knowledge Sources [0.0]
Multi-source multi-hop question answering (QA) represents a challenging task in natural language processing.
Existing methods often suffer from cascading errors, insufficient handling of knowledge conflicts, and computational inefficiency.
We propose Adaptive Multi-source Knowledge-Oriented Reasoning (AMKOR) to dynamically fuse parametric and retrieved knowledge.
arXiv Detail & Related papers (2025-02-09T16:06:43Z) - BRiTE: Bootstrapping Reinforced Thinking Process to Enhance Language Model Reasoning [78.63421517563056]
Large Language Models (LLMs) have demonstrated remarkable capabilities in complex reasoning tasks.
We present a unified probabilistic framework that formalizes LLM reasoning through a novel graphical model.
We introduce the Bootstrapping Reinforced Thinking Process (BRiTE) algorithm, which works in two steps.
arXiv Detail & Related papers (2025-01-31T02:39:07Z) - Insight-V: Exploring Long-Chain Visual Reasoning with Multimodal Large Language Models [64.1799100754406]
Large Language Models (LLMs) demonstrate enhanced capabilities and reliability by reasoning more.
Despite various efforts to improve LLM reasoning, high-quality long-chain reasoning data and optimized training pipelines still remain inadequately explored in vision-language tasks.
We present Insight-V, an early effort to 1) scalably produce long and robust reasoning data for complex multi-modal tasks, and 2) an effective training pipeline to enhance the reasoning capabilities of MLLMs.
arXiv Detail & Related papers (2024-11-21T18:59:55Z) - Q-SFT: Q-Learning for Language Models via Supervised Fine-Tuning [62.984693936073974]
Value-based reinforcement learning can learn effective policies for a wide range of multi-turn problems.
Current value-based RL methods have proven particularly challenging to scale to the setting of large language models.
We propose a novel offline RL algorithm that addresses these drawbacks, casting Q-learning as a modified supervised fine-tuning problem.
arXiv Detail & Related papers (2024-11-07T21:36:52Z) - Plan of Thoughts: Heuristic-Guided Problem Solving with Large Language Models [0.0]
We formalize a planning-based approach to perform multi-step problem solving with language models.
We demonstrate a superior success rate of 89.4% on the Game of 24 task as compared to existing approaches.
arXiv Detail & Related papers (2024-04-29T18:51:17Z) - Distilling Algorithmic Reasoning from LLMs via Explaining Solution Programs [2.3020018305241337]
Distilling explicit chain-of-thought reasoning paths has emerged as an effective method for improving the reasoning abilities of large language models.
We propose a novel approach to distill reasoning abilities from LLMs by leveraging their capacity to explain solutions.
Our experiments demonstrate that learning from explanations enables the Reasoner to more effectively guide program implementation by a Coder.
arXiv Detail & Related papers (2024-04-11T22:19:50Z) - Entropy-Regularized Token-Level Policy Optimization for Language Agent Reinforcement [67.1393112206885]
Large Language Models (LLMs) have shown promise as intelligent agents in interactive decision-making tasks.
We introduce Entropy-Regularized Token-level Policy Optimization (ETPO), an entropy-augmented RL method tailored for optimizing LLMs at the token level.
We assess the effectiveness of ETPO within a simulated environment that models data science code generation as a series of multi-step interactive tasks.
arXiv Detail & Related papers (2024-02-09T07:45:26Z) - DialCoT Meets PPO: Decomposing and Exploring Reasoning Paths in Smaller
Language Models [18.96271708412086]
Chain-of-Thought (CoT) prompting has proven to be effective in enhancing the reasoning capabilities of Large Language Models (LLMs) with at least 100 billion parameters.
We introduce Dialogue-guided Chain-of-Thought (DialCoT) which employs a dialogue format to generate intermediate reasoning steps, guiding the model toward the final answer.
arXiv Detail & Related papers (2023-10-08T08:52:13Z) - KECP: Knowledge Enhanced Contrastive Prompting for Few-shot Extractive
Question Answering [28.18555591429343]
We propose a novel framework named Knowledge Enhanced Contrastive Prompt-tuning (KECP)
Instead of adding pointer heads to PLMs, we transform the task into a non-autoregressive Masked Language Modeling (MLM) generation problem.
Our method consistently outperforms state-of-the-art approaches in few-shot settings by a large margin.
arXiv Detail & Related papers (2022-05-06T08:31:02Z) - Efficient Model-Based Multi-Agent Mean-Field Reinforcement Learning [89.31889875864599]
We propose an efficient model-based reinforcement learning algorithm for learning in multi-agent systems.
Our main theoretical contributions are the first general regret bounds for model-based reinforcement learning for MFC.
We provide a practical parametrization of the core optimization problem.
arXiv Detail & Related papers (2021-07-08T18:01:02Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.