Visual-OPSD: Cross-Modal On-Policy Self-Distillation for Efficient Unified Multimodal Reasoning

arXiv:2606.18974 · cs.CV · Submitted 2026-06-17 · Read on arXiv

cs.CV

Submitted: 2026-06-17

Updated: 2026-09-25

Code: https://github.com/TiezMind/Visual-OPSD

Terminology

• Distillation — 46× in this paper · explained in 2 episodes
• Benchmark — 32× in this paper · explained in 2 episodes
• Accuracy — 22× in this paper · explained in 1 episode
• Self-distillation — 17× in this paper · explained in 6 episodes
• Trace — 16× in this paper · explained in 2 episodes
• On-Policy Self-Distillation — 13× in this paper
• Regularization — 5× in this paper · explained in 3 episodes
• Diffusion Model — 3× in this paper · explained in 8 episodes
• Multimodal Reasoning — 3× in this paper · explained in 5 episodes
• Reasoning Traces — 3× in this paper · explained in 3 episodes
• Alignment — 2× in this paper · explained in 1 episode
• Fine-Tuning — 2× in this paper · explained in 5 episodes
• Grounding — 2× in this paper · explained in 3 episodes
• Knowledge Distillation — 2× in this paper · explained in 11 episodes
• On-policy Self-Distillation (OPSD) — 2× in this paper · explained in 2 episodes
• Reinforcement Learning — 2× in this paper · explained in 4 episodes
• Vision-Language Models (VLMs) — 2× in this paper · explained in 6 episodes
• Activation Checkpointing — 1× in this paper · explained in 1 episode
• Core — 1× in this paper · explained in 2 episodes
• Decoupling — 1× in this paper · explained in 2 episodes
• Failure Modes — 1× in this paper · explained in 3 episodes
• Foundation Model — 1× in this paper · explained in 3 episodes
• Foundation Models — 1× in this paper · explained in 7 episodes
• Generalization — 1× in this paper · explained in 13 episodes
• Information Asymmetry — 1× in this paper · explained in 2 episodes
• KV cache — 1× in this paper · explained in 3 episodes
• Robustness — 1× in this paper · explained in 5 episodes
• Transformer — 1× in this paper · explained in 2 episodes
• Verification — 1× in this paper · explained in 2 episodes

Sources

• Qwen3-VL Technical Report
• Chameleon: Mixed-Modal Early-Fusion Foundation Models
• Janus-Pro: Unified Multimodal Understanding and Generation with Data and Model Scaling
• Emerging Properties in Unified Multimodal Pretraining
• BLINK: Multimodal Large Language Models Can See but Not Perceive
• Distilling the Knowledge in a Neural Network
• Visual Sketchpad: Sketching as a Visual Chain of Thought for Multimodal Language Models
• ChartQA: A Benchmark for Question Answering about Charts with Visual and Logical Reasoning
• SAT: Dynamic Spatial Aptitude Training for Multimodal Language Models
• Visual Chain of Thought: Bridging Logical Gaps with Multimodal Infillings
• Cambrian-1: A Fully Open, Vision-Centric Exploration of Multimodal LLMs
• Eyes Wide Shut? Exploring the Visual Shortcomings of Multimodal LLMs
• InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiency
• Emu3: Next-Token Prediction is All You Need
• Zooming without Zooming: Region-to-Image Distillation for Fine-Grained Multimodal Perception
• Janus: Decoupling Visual Encoding for Unified Multimodal Understanding and Generation
• VSP: Assessing the dual challenges of perception and reasoning in spatial planning tasks for VLMs
• Vision-OPD: Learning to See Fine Details for Multimodal LLMs via On-Policy Self-Distillation
• OPSDL: On-Policy Self-Distillation for Long-Context Language Models
• Multimodal Chain-of-Thought Reasoning in Language Models

Related papers

• Loss Knows Best: Detecting Annotation Errors in Videos via Loss Trajectories
• AnchorWeave: World-Consistent Video Generation with Retrieved Local Spatial Memories
• Benchmarking the Robustness of Foundation Models for Mammography under Domain Shift
• MambaX-Net: Dual-Input Mamba-Enhanced Cross-Attention Network for Longitudinal MRI Segmentation
• TeleOCR: Navigating Document Parsing Across Digital and Camera-Captured Documents
• A Survey on Efficient Vision-Language-Action Models