# Reasoning as Representation: Rethinking Visual Reinforcement Learning in Image Quality Assessment

arXiv: https://arxiv.org/abs/2510.11369 · HTML: https://mlsanity.com/paper/2510.11369/

## Key facts

- Cited by 18 papers in the map (data through 14 Oct 2026).
- Placed in Multimodal reasoning, one of 17 directions in LLM reasoning.
- Builds on 17 of its 77 references that are in the map.
- First posted on arXiv 13 Oct 2025.

## Placed in

- LLM reasoning / [Multimodal reasoning](https://mlsanity.com/lr/multimodal-reasoning/)

## Builds on (most cited first, 17 of 17)

- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [Qwen2.5-VL Technical Report](https://mlsanity.com/paper/2502.13923/) (2025)
- [Decoupled Weight Decay Regularization](https://mlsanity.com/paper/1711.05101/) (2017)
- [Qwen2.5-Coder Technical Report](https://mlsanity.com/paper/2409.12186/) (2024)
- [Qwen2.5-Math Technical Report: Toward Mathematical Expert Model via Self-Improvement](https://mlsanity.com/paper/2409.12122/) (2024)
- [Flow-GRPO: Training Flow Matching Models via Online RL](https://mlsanity.com/paper/2505.05470/) (2025)
- [Visual-RFT: Visual Reinforcement Fine-Tuning](https://mlsanity.com/paper/2503.01785/) (2025)
- [MedVLM-R1: Incentivizing Medical Reasoning Capability of Vision-Language Models (VLMs) via Reinforcement Learning](https://mlsanity.com/paper/2502.19634/) (2025)
- [InternLM-Math: Open Math Large Language Models Toward Verifiable Reasoning](https://mlsanity.com/paper/2402.06332/) (2024)
- [Unified Multimodal Chain-of-Thought Reward Model through Reinforcement Fine-Tuning](https://mlsanity.com/paper/2505.03318/) (2025)
- [o1-Coder: an o1 Replication for Coding](https://mlsanity.com/paper/2412.00154/) (2024)
- [Enhancing Descriptive Image Quality Assessment with A Large-scale Multi-modal Dataset](https://mlsanity.com/paper/2405.18842/) (2024)
- [Q-Eval-100K: Evaluating Visual Quality and Alignment Level for Text-to-Vision Content](https://mlsanity.com/paper/2503.02357/) (2025)
- [VQ-Insight: Teaching VLMs for AI-Generated Video Quality Understanding via Progressive Visual Reinforcement Learning](https://mlsanity.com/paper/2506.18564/) (2025)
- [Q-SiT: Teaching LMMs for Image Quality Scoring and Interpreting](https://mlsanity.com/paper/2503.09197/) (2025)
- [AvatarShield: Visual Reinforcement Learning for Human-Centric Synthetic Video Detection](https://mlsanity.com/paper/2505.15173/) (2025)

## Built on by (most cited first, 18 of 18)

- [AutoSOTA: An End-to-End Automated Research System for State-of-the-Art AI Model Discovery](https://mlsanity.com/paper/2604.05550/) (2026)
- [Zoom-IQA: Image Quality Assessment with Reliable Region-Aware Reasoning](https://mlsanity.com/paper/2601.02918/) (2026)
- [UARE: A Unified Vision-Language Model for Image Quality Assessment, Restoration, and Enhancement](https://mlsanity.com/paper/2512.06750/) (2025)
- [TalkPhoto: A Versatile Training-Free Conversational Assistant for Intelligent Image Editing](https://mlsanity.com/paper/2601.01915/) (2026)
- [Q-Probe: Scaling Image Quality Assessment to High Resolution via Context-Aware Agentic Probing](https://mlsanity.com/paper/2601.15356/) (2026)
- [OARS: Process-Aware Online Alignment for Generative Real-World Image Super-Resolution](https://mlsanity.com/paper/2603.12811/) (2026)
- [One Model, Two Minds: Task-Conditioned Reasoning for Unified Image Quality and Aesthetic Assessment](https://mlsanity.com/paper/2603.19779/) (2026)
- [Banana100: Breaking NR-IQA Metrics by 100 Iterative Image Replications with Nano Banana Pro](https://mlsanity.com/paper/2604.03400/) (2026)
- [Beyond Time Shifts: Adapting Omni-LLM as a Reference-Free Evaluator for Generative Audio-Visual Models](https://mlsanity.com/paper/2607.09091/) (2026)
- [RewardVerse: Rubric-Guided Policy Optimization for Video Reward Modeling](https://mlsanity.com/paper/2609.22947/) (2026)
- [Q-Tacit: Image Quality Assessment via Latent Visual Reasoning](https://mlsanity.com/paper/2603.22641/) (2026)
- [VQ-Jarvis: Retrieval-Augmented Video Restoration Agent with Sharp Vision and Fast Thought](https://mlsanity.com/paper/2603.22998/) (2026)
- [Redefining Quality Criteria and Distance-Aware Score Modeling for Image Editing Assessment](https://mlsanity.com/paper/2604.12175/) (2026)
- [VersusQ: Pairwise Margin Reasoning for Generalizable Video Quality Assessment](https://mlsanity.com/paper/2605.21130/) (2026)
- [MR-IQA: A Unified Margin View of Regression and Ranking for Blind Image Quality Assessment](https://mlsanity.com/paper/2606.29760/) (2026)
- [SpatialQ: Understanding 3D Gaussian Splatting Scene Quality via Visual-based MLLM](https://mlsanity.com/paper/2607.26595/) (2026)
- [Bridging Adversarial and Collaborative Learning for AI-Generated Image Quality Assessment](https://mlsanity.com/paper/2608.24372/) (2026)
- [Learning from Failure: Leveraging Unreliable Predictions in Semi-Supervised Real-World Adverse Weather Removal](https://mlsanity.com/paper/2610.02051/) (2026)

## Abstract (arXiv)

Reasoning-based image quality assessment (IQA) models trained through reinforcement learning (RL) exhibit exceptional generalization, yet the underlying mechanisms and critical factors driving this capability remain underexplored in current research. Moreover, despite their superior performance, these models incur inference energy usage and latency orders of magnitude higher than their earlier counterparts, restricting their deployment in specific scenarios. Through extensive experiments, this paper verifies and elaborates that through RL training, MLLMs leverage their reasoning capability to convert redundant visual representations into compact, cross-domain aligned text representations. This conversion is precisely the source of the generalization exhibited by these reasoning-based IQA models. Building on this fundamental insight, we propose a novel algorithm, RALI, which employs contrastive learning to directly align images with these generalizable text representations learned by RL. This approach eliminates the reliance on reasoning processes and even obviates the need to load an LLM. For the quality scoring task, this framework achieves generalization performance comparable to reasoning-based models while requiring less than 5% of their model parameters and inference time.