mlsanityData to 7 Oct 2026

VisualDeltas: Learning Preferences from Visual Quality Perturbations

arXiv 2603.07272 ↗First posted 7 Mar 2026cs.AIcited by 0

Not on the map. Why?

Abstract (arXiv)

We present VisualDeltas, a lightweight preference-learning framework that extracts supervision from visual quality variations in multimodal data. By leveraging the systematic impact of image quality on visual perception and reasoning, VisualDeltas induces informative preference signals without relying on human annotations or external teachers. The framework supports both label-free and label-based regimes, enabling flexible use of available supervision when present. Across diverse multimodal benchmarks and model scales, VisualDeltas consistently outperforms rejection-sampling fine-tuning and improves generalization, and extends naturally to a range of visual degradations.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 19 of its 26 references that are in the map. First posted on arXiv 7 Mar 2026.

Builds on

19 of 26 references.

References · 19

  1. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 7,916
  2. 2022Constitutional AI: Harmlessness from AI Feedbackcited by 2,950
  3. 2016Making the V in VQA Matter: Elevating the Role of Image Understanding in Visual Question Answeringcited by 2,498
  4. 2024Self-Rewarding Language Modelscited by 654
  5. 2024Measuring Multimodal Mathematical Reasoning with MATH-Vision Datasetcited by 516
  6. 2015Compositional Semantic Parsing on Semi-Structured Tablescited by 509
  7. 2024Direct Preference Optimization of Video Large Multimodal Models from Language Model Rewardcited by 145
  8. 2025MM-RLHF: The Next Step Forward in Multimodal LLM Alignmentcited by 79
  9. 2024Relative Preference Optimization: Enhancing LLM Alignment through Contrasting Responses across Identical and Diverse Promptscited by 36
  10. 2024Modality-Fair Preference Optimization for Trustworthy MLLM Alignmentcited by 26
  11. 2025The Delta Learning Hypothesis: Preference Tuning on Weak Data can Yield Strong Gainscited by 25
  12. 2024MallowsPO: Fine-Tune Your LLM with Preference Dispersionscited by 18
7 more
  1. 2025Mitigating Hallucinations in Large Vision-Language Models via Entity-Centric Multimodal Preference Optimizationcited by 10
  2. 2025Robust-R1: Degradation-Aware Reasoning for Robust Visual Understandingcited by 7
  3. 2025AdaViP: Aligning Multi-modal LLMs via Adaptive Vision-enhanced Preference Optimizationcited by 6
  4. 2024MutualVPR: A Mutual Learning Framework for Resolving Supervision Inconsistencies via Adaptive Clusteringcited by 5
  5. 2025LPOI: Listwise Preference Optimization for Vision Language Modelscited by 5
  6. 2025M3PO: Multimodal-Model-Guided Preference Optimization for Visual Instruction Followingcited by 1
  7. 2025Robust Diagram Reasoning: A Framework for Enhancing LVLM Performance on Visually Perturbed Scientific Diagramscited by 1

Built on by

Not cited yet.