SATORI-R1: Incentivizing Multimodal Reasoning with Spatial Grounding and Verifiable Rewards

• URL: http://arxiv.org/abs/2505.19094v1
• Date: Sun, 25 May 2025 11:11:06 GMT
• Title: SATORI-R1: Incentivizing Multimodal Reasoning with Spatial Grounding and Verifiable Rewards
• Authors: Chuming Shen, Wei Wei, Xiaoye Qu, Yu Cheng,
• Abstract summary: DeepSeek-R1 has demonstrated powerful reasoning capabilities in the text domain through stable reinforcement learning (RL)<n>In this paper, we introduce SATORI ($textbfSpatially$ $textbfAnchored$ $textbfTask$ $textbfOptimization$ with $textbfRetextbfInforcement$ Learning), which decomposes VQA into three verifiable stages.<n> Experiments demonstrate consistent performance improvements across seven VQA benchmarks, achieving up to $15.7%$ improvement in
• Score: 23.02076024811612
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: DeepSeek-R1 has demonstrated powerful reasoning capabilities in the text domain through stable reinforcement learning (RL). Recently, in the multimodal domain, works have begun to directly apply RL to generate R1-like free-form reasoning for Visual Question Answering (VQA) tasks. However, multimodal tasks share an intrinsically different nature from textual tasks, which heavily rely on the understanding of the input image to solve the problem. Therefore, such free-form reasoning faces two critical limitations in the VQA task: (1) Extended reasoning chains diffuse visual focus away from task-critical regions, degrading answer accuracy. (2) Unverifiable intermediate steps amplify policy-gradient variance and computational costs overhead. To address these issues, in this paper, we introduce SATORI ($\textbf{S}patially$ $\textbf{A}nchored$ $\textbf{T}ask$ $\textbf{O}ptimization$ with $\textbf{R}e\textbf{I}nforcement$ Learning), which decomposes VQA into three verifiable stages, including global image captioning, region localization, and answer prediction, each supplying explicit reward signals. Furthermore, we also introduce VQA-Verify, a 12k dataset annotated with answer-aligned captions and bounding-boxes to facilitate training. Experiments demonstrate consistent performance improvements across seven VQA benchmarks, achieving up to $15.7\%$ improvement in accuracy in accuracy compared to the R1-like baseline. Our analysis of the attention map confirms enhanced focus on critical regions, which brings improvements in accuracy. Our code is available at https://github.com/justairr/SATORI-R1.

Related papers

• From Verifiable Dot to Reward Chain: Harnessing Verifiable Reference-based Rewards for Reinforcement Learning of Open-ended Generation [52.62655622099456]
  We propose reinforcement learning with verifiable reference-based rewards (RLVRR)<n>Instead of checking the final answer, RLVRR extracts an ordered linguistic signal from high-quality references (i.e., reward chain)<n>In this way, RLVRR decomposes rewards into two dimensions: content, which preserves deterministic core concepts, and style, which evaluates adherence to stylistic properties.
  arXiv  Detail & Related papers  (2026-01-26T14:39:58Z)
• Eigen-1: Adaptive Multi-Agent Refinement with Monitor-Based RAG for Scientific Reasoning [53.45095336430027]
  We develop a unified framework that combines implicit retrieval and structured collaboration.<n>On Humanity's Last Exam (HLE) Bio/Chem Gold, our framework achieves 48.3% accuracy.<n>Results on SuperGPQA and TRQA confirm robustness across domains.
  arXiv  Detail & Related papers  (2025-09-25T14:05:55Z)
• From Sufficiency to Reflection: Reinforcement-Guided Thinking Quality in Retrieval-Augmented Reasoning for LLMs [13.410543801811992]
  This paper analyzes existing RAG reasoning models and identifies three main failure patterns.<n>We propose TIRESRAG-R1, a novel framework using a think-retrieve-reflect process and a multi-dimensional reward system.<n>Experiments on four multi-hop QA datasets show that TIRESRAG-R1 outperforms prior RAG methods and generalizes well to single-hop tasks.
  arXiv  Detail & Related papers  (2025-07-30T14:29:44Z)
• SVQA-R1: Reinforcing Spatial Reasoning in MLLMs via View-Consistent Reward Optimization [57.484274282231226]
  We propose SVQA-R1, the first framework to extend R1-style training to spatial VQA.<n>In particular, we introduce Spatial-GRPO, a novel group-wise RL strategy that constructs view-consistent rewards by perturbing spatial relations between objects.<n>Our model, SVQA-R1, not only dramatically improved accuracy on spatial VQA benchmarks but also exhibits interpretable reasoning paths even without using supervised fine-tuning data.
  arXiv  Detail & Related papers  (2025-06-02T06:58:43Z)
• Reinforcing Video Reasoning with Focused Thinking [65.85683941058916]
  We propose TW-GRPO, a novel framework that enhances visual reasoning with focused thinking and dense reward granularity.<n>Specifically, we employ a token weighting mechanism that prioritizes tokens with high informational density.<n>We also reformulate RL training by shifting from single-choice to multi-choice QA tasks.
  arXiv  Detail & Related papers  (2025-05-30T15:42:19Z)
• VLM-R$^3$: Region Recognition, Reasoning, and Refinement for Enhanced Multimodal Chain-of-Thought [51.43082554363725]
  We introduce textbfVLM-R$3$ (textbfVisual textbfLanguage textbfModel with textbfRegion textbfRecognition and textbfReasoning), a framework that equips an MLLM with the ability to decide emph when additional visual evidence is needed.<n>Experiments on MathVista, ScienceQA, and other benchmarks show that VLM-R$3$ sets a new
  arXiv  Detail & Related papers  (2025-05-22T03:50:13Z)
• From Head to Tail: Towards Balanced Representation in Large Vision-Language Models through Adaptive Data Calibration [30.781359402734036]
  Large Vision-Language Models (LVLMs) have achieved significant progress in combining visual comprehension with language generation.<n>Despite this success, the training data of LVLMs still suffers from Long-Tail (LT) problems, where the data distribution is highly imbalanced.<n>We propose an $textbfA$daptive $textbfD$ata $textbfR$ebalancing, while in the DS stage, we leverage Denoising Diffusion Probabilistic Models (DDPMs) and scarce images to supplement underrepresented portions.
  arXiv  Detail & Related papers  (2025-03-17T05:01:09Z)
• Visual-RFT: Visual Reinforcement Fine-Tuning [75.20572976629646]
  Reinforcement Fine-Tuning (RFT) in Large Reasoning Models like OpenAI o1 learns from feedback on its answers.<n>Visual-RFT further extends the application areas of RFT on visual tasks.
  arXiv  Detail & Related papers  (2025-03-03T18:16:32Z)
• FLARE: Faithful Logic-Aided Reasoning and Exploration [50.9814063216852]
  We introduce a novel approach for traversing the problem space using task decompositions.<n>We use the Large Language Models to plan a solution, soft-formalise the query into facts and predicates using a logic programming code.<n>Our method allows us to compute the faithfulness of the reasoning process w.r.t. the generated code and analyse the steps of the multi-hop search without relying on external solvers.
  arXiv  Detail & Related papers  (2024-10-14T19:39:11Z)
• Can We Verify Step by Step for Incorrect Answer Detection? [22.984011562264147]
  We introduce a benchmark, R2PE, designed specifically to explore the relationship between reasoning chains and performance in various reasoning tasks.<n>This benchmark aims to measure the falsehood of the final output of LLMs based on the reasoning steps.<n>We propose the process discernibility score (PDS) framework that beats the answer-checking baseline by a large margin.
  arXiv  Detail & Related papers  (2024-02-16T09:29:50Z)
• Exploring Question Decomposition for Zero-Shot VQA [99.32466439254821]
  We investigate a question decomposition strategy for visual question answering. We show that naive application of model-written decompositions can hurt performance. We introduce a model-driven selective decomposition approach for second-guessing predictions and correcting errors.
  arXiv  Detail & Related papers  (2023-10-25T23:23:57Z)
• Modularized Zero-shot VQA with Pre-trained Models [20.674979268279728]
  We propose a modularized zero-shot network that explicitly decomposes questions into sub reasoning steps and is highly interpretable. Our experiments on two VQA benchmarks under the zero-shot setting demonstrate the effectiveness of our method.
  arXiv  Detail & Related papers  (2023-05-27T05:00:14Z)
• Toward Unsupervised Realistic Visual Question Answering [70.67698100148414]
  We study the problem of realistic VQA (RVQA), where a model has to reject unanswerable questions (UQs) and answer answerable ones (AQs) We first point out 2 drawbacks in current RVQA research, where (1) datasets contain too many unchallenging UQs and (2) a large number of annotated UQs are required for training. We propose a new testing dataset, RGQA, which combines AQs from an existing VQA dataset with around 29K human-annotated UQs. This combines pseudo UQs obtained by randomly pairing images and questions, with an
  arXiv  Detail & Related papers  (2023-03-09T06:58:29Z)
• Structured Multimodal Attentions for TextVQA [57.71060302874151]
  We propose an end-to-end structured multimodal attention (SMA) neural network to mainly solve the first two issues above. SMA first uses a structural graph representation to encode the object-object, object-text and text-text relationships appearing in the image, and then designs a multimodal graph attention network to reason over it. Our proposed model outperforms the SoTA models on TextVQA dataset and two tasks of ST-VQA dataset among all models except pre-training based TAP.
  arXiv  Detail & Related papers  (2020-06-01T07:07:36Z)

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.