MIRROR: Learning from the Other View for Multi-Modal Reasoning

arXiv:2607.21552 · cs.AI, cs.LG · Submitted 2026-07-23 · Read on arXiv

cs.AI, cs.LG

Submitted: 2026-07-23

Updated: 2026-10-08

Terminology

• Benchmark — 1× in this paper · explained in 2 episodes
• Failure Modes — 1× in this paper · explained in 3 episodes
• Large Language Models (LLMs) — 1× in this paper · explained in 6 episodes
• Multimodal Reasoning — 1× in this paper · explained in 5 episodes
• Reinforcement Learning — 1× in this paper · explained in 4 episodes
• Vision-Language Models (VLMs) — 1× in this paper · explained in 6 episodes
• Vision-language models — 1× in this paper · explained in 2 episodes

Sources

• On-Policy Distillation of Language Models: Learning from Self-Generated Mistakes
• Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyond
• Qwen3-VL Technical Report
• VOLD: Reasoning Transfer from LLMs to Vision-Language Models via On-Policy Distillation
• OpenDataArena: A Fair and Open Arena for Benchmarking Post-Training Dataset Value
• Milestones over Outcome: Unlocking Geometric Reasoning with Sub-Goal Verifiable Reward
• Evaluating Large Language Models Trained on Code
• UnityMAS-O: A General RL Optimization Framework for LLM-Based Multi-Agent Systems
• InternVL: Scaling up Vision Foundation Models and Aligning for Generic Visual-Linguistic Tasks
• Insight-V: Exploring Long-Chain Visual Reasoning with Multimodal Large Language Models
• MiniLLM: On-Policy Distillation of Large Language Models
• Uni-OPD: Unifying On-Policy Distillation with a Dual-Perspective Recipe
• Vision-R1: Incentivizing Reasoning Capability in Multimodal Large Language Models
• GPT-4o System Card
• Reinforcement Learning via Self-Distillation
• Rethinking On-Policy Distillation of Large Language Models: Phenomenology, Mechanism, and Recipe
• A Survey of State of the Art Large Vision Language Models: Alignment, Benchmark, Evaluations and Challenges
• Continuous control with deep reinforcement learning
• Visual Instruction Tuning
• ChartQA: A Benchmark for Question Answering about Charts with Visual and Logical Reasoning

Related papers

• MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
• Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
• Agents' Last Exam
• The Clinician's Veto: Navigating Trust, Liability, and Uncertainty in Autonomous AI Prescribing
• MindHelper: Closed-Loop Embodied Mental-State Reasoning for Precision Intervention
• Incumbent Advantage: Brand Bias and Cognitive Manipulation Dynamics in LLM Recommendation Systems