Think or Not Think: A Study of Explicit Thinking in Rule-Based Visual Reinforcement Fine-Tuning

• URL: http://arxiv.org/abs/2503.16188v4
• Date: Mon, 12 May 2025 12:29:12 GMT
• Title: Think or Not Think: A Study of Explicit Thinking in Rule-Based Visual Reinforcement Fine-Tuning
• Authors: Ming Li, Jike Zhong, Shitian Zhao, Yuxiang Lai, Haoquan Zhang, Wang Bill Zhu, Kaipeng Zhang,
• Abstract summary: We first propose CLS-RL for MLLM image classification, using verifiable rewards for fine-tuning.<n>We then rethink and question whether explicit thinking in RFT is always necessary.<n>No-Thinking-RL explores RFT without thinking by introducing a simple equality accuracy reward.
• Score: 7.78764814568908
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: This paper investigates the role of explicit thinking process in rule-based reinforcement fine-tuning (RFT) for MLLMs. We first propose CLS-RL for MLLM image classification, using verifiable rewards for fine-tuning. Experiments show CLS-RL significantly outperforms SFT and yields a cross-dataset generalization effect. We then rethink and question whether explicit thinking in RFT is always necessary. Challenging the convention that explicit thinking is crucial for the success of RFT, we introduce No-Thinking-RL, exploring RFT without thinking by introducing a simple equality accuracy reward. We evaluate No-Thinking-RL on 6 diverse tasks across different model sizes and types. Experimental results reveal three key findings: 1). Visual perception tasks do not require thinking during RFT, as No-Thinking-RL consistently outperforms or matches Thinking-based RFT across model sizes. 2).} Models with limited capabilities struggle to generate high-quality CoT for RFT, making Thinking-based RFT less effective than No-Thinking-RL. 3). There are inconsistencies between the answers in the thinking and answer tags for some responses of thinking-based RFT, which show lower accuracy than the overall accuracy. We hypothesize that explicit thinking before verifiable answers may hinder reward convergence and reduce performance. To test this hypothesis, we propose Think-After-Answer, which places thinking after the answer to mitigate this effect for experimental verification. Lastly, we conduct a pilot study to explore whether MLLMs can learn when to think during RFT, introducing an Adaptive-Thinking method. Experiments show that it converges to a specific prompt depending on model capability and task complexity, achieving comparable or better performance than both Thinking and No-Thinking-RL. This suggests MLLMs can adaptively decide to think or not based on their capabilities and task complexity.

Related papers

• RLPR: Extrapolating RLVR to General Domains without Verifiers [103.14103272635893]
  We propose RLPR, a simple verifier-free framework that extrapolates RLVR to broader general domains.<n>We find that addressing the high variance of this noisy probability reward is crucial to make it work.<n>RLPR consistently improves reasoning capabilities in both areas for Gemma, Llama, and Qwen based models.
  arXiv  Detail & Related papers  (2025-06-23T02:56:36Z)
• R3-RAG: Learning Step-by-Step Reasoning and Retrieval for LLMs via Reinforcement Learning [62.742230250513025]
  Retrieval-Augmented Generation (RAG) integrates external knowledge with Large Language Models (LLMs) to enhance factual correctness and hallucination.<n>We propose $textbfR3-RAG$, which uses $textbfR$einforcement learning to make the LLM learn how to $textbfR$eason and $textbfR$etrieve step by step, thus retrieving comprehensive external knowledge and leading to correct answers.
  arXiv  Detail & Related papers  (2025-05-26T12:25:37Z)
• Thinkless: LLM Learns When to Think [57.857534644932194]
  Reasoning Language Models, capable of extended chain-of-thought reasoning, have demonstrated remarkable performance on tasks requiring complex logical inference.<n>We propose Thinkless, a learnable framework that empowers an LLM to adaptively select between short-form and long-form reasoning.<n>On several benchmarks such as Minerva Algebra, MATH-500, and GSM8K, Thinkless is able to reduce the usage of long-chain thinking by 50% - 90%.
  arXiv  Detail & Related papers  (2025-05-19T17:24:16Z)
• Learning When to Think: Shaping Adaptive Reasoning in R1-Style Models via Multi-Stage RL [19.731871225975926]
  Large reasoning models (LRMs) are proficient at generating explicit, step-by-step reasoning sequences before producing final answers.<n>To address this over-thinking problem, we explore how to equip LRMs with adaptive thinking capabilities.<n>We propose AutoThink, a multi-stage reinforcement learning framework that progressively optimize reasoning policies.
  arXiv  Detail & Related papers  (2025-05-16T04:01:57Z)
• VL-Rethinker: Incentivizing Self-Reflection of Vision-Language Models with Reinforcement Learning [55.97950660659051]
  GPT-o1 and DeepSeek-R1 have demonstrated great potential in solving challenging problems through explicit reflection. In this paper, we aim to enhance the slow-thinking capabilities of vision-language models using reinforcement learning. Our model, VL-Rethinker, advances state-of-the-art scores on MathVista, MathVerse, and MathVision to achieve 80.3%, 61.8%, and 43.9% respectively.
  arXiv  Detail & Related papers  (2025-04-10T17:41:56Z)
• On the Suitability of Reinforcement Fine-Tuning to Visual Tasks [15.971601297360227]
  Researchers have been starting to apply RFT to MLLMs, hoping it will also enhance the capabilities of visual understanding. In this work, we endeavor to understand the suitabilities and limitations of RFT for visual tasks, through experimental analysis and observations.
  arXiv  Detail & Related papers  (2025-04-08T04:45:00Z)
• Exploring the Effect of Reinforcement Learning on Video Understanding: Insights from SEED-Bench-R1 [53.894789613838654]
  We introduce SEED-Bench-R1, a benchmark designed to evaluate post-training methods for MLLMs in video understanding. It includes intricate real-world videos and complex everyday planning tasks in the format of multiple-choice questions. Using Qwen2-VL-Instruct-7B as a base model, we compare RL with supervised fine-tuning (SFT) Our detailed analysis reveals that RL enhances visual perception but often produces less coherent reasoning chains.
  arXiv  Detail & Related papers  (2025-03-31T17:55:23Z)
• Innate Reasoning is Not Enough: In-Context Learning Enhances Reasoning Large Language Models with Less Overthinking [39.48406368755411]
  Large Language Models (LLMs) have introduced Reasoning Large Language Models (RLLMs) RLLMs exhibit innate Chain-of-Thought (CoT) reasoning capability obtained from training, leading to a natural question: "Is CoT prompting necessary to enhance the reasoning capability of RLLMs?" We present the first comprehensive analysis of the impacts of Zero-shot CoT and Few-shot CoT on RLLMs across mathematical reasoning tasks.
  arXiv  Detail & Related papers  (2025-03-25T12:37:22Z)
• OpenVLThinker: An Early Exploration to Complex Vision-Language Reasoning via Iterative Self-Improvement [91.88062410741833]
  This study investigates whether similar reasoning capabilities can be successfully integrated into large vision-language models (LVLMs) We consider an approach that iteratively leverages supervised fine-tuning (SFT) on lightweight training data and Reinforcement Learning (RL) to further improve model generalization. OpenVLThinker, a LVLM exhibiting consistently improved reasoning performance on challenging benchmarks such as MathVista, MathVerse, and MathVision, demonstrates the potential of our strategy for robust vision-language reasoning.
  arXiv  Detail & Related papers  (2025-03-21T17:52:43Z)
• GTR: Guided Thought Reinforcement Prevents Thought Collapse in RL-based VLM Agent Training [62.536191233049614]
  Reinforcement learning with verifiable outcome rewards (RLVR) has effectively scaled up chain-of-thought (CoT) reasoning in large language models (LLMs) This work investigates this problem through extensive experiments on complex card games, such as 24 points, and embodied tasks from ALFWorld. We find that when rewards are based solely on action outcomes, RL fails to incentivize CoT reasoning in VLMs, instead leading to a phenomenon we termed thought collapse.
  arXiv  Detail & Related papers  (2025-03-11T15:17:02Z)
• LIMR: Less is More for RL Scaling [25.477841726836836]
  We introduce Learning Impact Measurement (LIM), an automated method to evaluate and prioritize training samples.<n>Our method achieves comparable or even superior performance using only 1,389 samples versus the full 8,523 samples dataset.<n>For reproducible research and future innovation, we are open-sourcing LIMR, including implementation of LIM, training and evaluation code, curated datasets, and trained models.
  arXiv  Detail & Related papers  (2025-02-17T15:13:29Z)
• On the Emergence of Thinking in LLMs I: Searching for the Right Intuition [34.32871896067864]
  We propose a post-training framework called Reinforcement Learning via Self-Play (RLSP)<n> RLSP involves three steps: supervised fine-tuning with human or synthetic demonstrations of the reasoning process, using an exploration reward signal to encourage diverse and efficient reasoning behaviors, and RL training with an outcome verifier to ensure correctness while preventing reward hacking.<n> Empirical studies in the math domain show that RLSP improves reasoning.
  arXiv  Detail & Related papers  (2025-02-10T18:52:04Z)
• SD-LoRA: Scalable Decoupled Low-Rank Adaptation for Class Incremental Learning [73.93639228235622]
  Continual Learning with foundation models has emerged as a promising paradigm to exploit abundant knowledge acquired during pre-training for tackling sequential tasks.<n>Existing prompt-based and Low-Rank Adaptation-based (LoRA-based) methods often require expanding a prompt/LoRA pool or retaining samples of previous tasks.<n>We propose Scalable Decoupled LoRA (SD-LoRA) for class incremental learning, which continually separates the learning of the magnitude and direction of LoRA components without rehearsal.
  arXiv  Detail & Related papers  (2025-01-22T20:00:41Z)
• Do NOT Think That Much for 2+3=? On the Overthinking of o1-Like LLMs [76.43407125275202]
  o1-like models can emulate human-like long-time thinking during inference. This paper presents the first comprehensive study on the prevalent issue of overthinking in these models. We propose strategies to mitigate overthinking, streamlining reasoning processes without compromising accuracy.
  arXiv  Detail & Related papers  (2024-12-30T18:55:12Z)
• Improving Large Language Models via Fine-grained Reinforcement Learning with Minimum Editing Constraint [104.53687944498155]
  Reinforcement learning (RL) has been widely used in training large language models (LLMs) We propose a new RL method named RLMEC that incorporates a generative model as the reward model. Based on the generative reward model, we design the token-level RL objective for training and an imitation-based regularization for stabilizing RL process.
  arXiv  Detail & Related papers  (2024-01-11T17:58:41Z)
• The Languini Kitchen: Enabling Language Modelling Research at Different Scales of Compute [66.84421705029624]
  We introduce an experimental protocol that enables model comparisons based on equivalent compute, measured in accelerator hours. We pre-process an existing large, diverse, and high-quality dataset of books that surpasses existing academic benchmarks in quality, diversity, and document length. This work also provides two baseline models: a feed-forward model derived from the GPT-2 architecture and a recurrent model in the form of a novel LSTM with ten-fold throughput.
  arXiv  Detail & Related papers  (2023-09-20T10:31:17Z)
• To Repeat or Not To Repeat: Insights from Scaling LLM under Token-Crisis [50.31589712761807]
  Large language models (LLMs) are notoriously token-hungry during pre-training, and high-quality text data on the web is approaching its scaling limit for LLMs. We investigate the consequences of repeating pre-training data, revealing that the model is susceptible to overfitting. Second, we examine the key factors contributing to multi-epoch degradation, finding that significant factors include dataset size, model parameters, and training objectives.
  arXiv  Detail & Related papers  (2023-05-22T17:02:15Z)
• SLCA: Slow Learner with Classifier Alignment for Continual Learning on a Pre-trained Model [73.80068155830708]
  We present an extensive analysis for continual learning on a pre-trained model (CLPM) We propose a simple but extremely effective approach named Slow Learner with Alignment (SLCA) Across a variety of scenarios, our proposal provides substantial improvements for CLPM.
  arXiv  Detail & Related papers  (2023-03-09T08:57:01Z)
• Efficient Gaussian Process Model on Class-Imbalanced Datasets for Generalized Zero-Shot Learning [37.00463358780726]
  We propose a Neural Network model that learns a latent feature embedding and a Gaussian Process (GP) regression model that predicts latent feature prototypes of unseen classes. Our model is trained efficiently with a simple training strategy that mitigates the impact of class-imbalanced training data.
  arXiv  Detail & Related papers  (2022-10-11T04:57:20Z)
• Self-Supervised Class Incremental Learning [51.62542103481908]
  Existing Class Incremental Learning (CIL) methods are based on a supervised classification framework sensitive to data labels. When updating them based on the new class data, they suffer from catastrophic forgetting: the model cannot discern old class data clearly from the new. In this paper, we explore the performance of Self-Supervised representation learning in Class Incremental Learning (SSCIL) for the first time.
  arXiv  Detail & Related papers  (2021-11-18T06:58:19Z)
• CRL: Class Representative Learning for Image Classification [5.11566193457943]
  We propose a novel model, called Class Representative Learning Model (CRL), that can be especially effective in image classification influenced by ZSL. In the CRL model, first, the learning step is to build class representatives to represent classes in datasets by aggregating prominent features extracted from a Convolutional Neural Network (CNN) The proposed CRL model demonstrated superior performance compared to the current state-of-the-art research in ZSL and mobile deep learning.
  arXiv  Detail & Related papers  (2020-02-16T17:02:59Z)

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.