Self-Supervised Visual On-Policy Distillation

Hugging Face Daily Papers Papers

Summary

The paper introduces S^2VOPD, a self-supervised method that improves vision-language models by distilling from original images into strongly augmented student views, achieving performance surpassing GPT-5.4 on benchmarks.

Visual on-policy distillation relies heavily on an informative teacher-student asymmetry, through either a larger, stronger teacher or privileged supervision, such as reference answers or ground-truth regions of interest. This raises a fundamental question: where can informative asymmetry come from when nothing privileged is available? We answer this by inverting where the asymmetry comes from. Rather than adding privileged information to the teacher, we subtract information from the student. This asymmetry creates the same effective learning signal for free as a teacher with access to information unavailable to the student, without ground-truth annotations, rewards, or a separate stronger teacher model. Building on this principle, we introduce Self-Supervised Visual On-Policy Distillation (S^2VOPD), a simple yet effective method that constructs on-policy learning signals from asymmetric augmented views. S^2VOPD distills the teacher's distribution conditioned on the original image on-policy into the student distribution conditioned on a strongly augmented view of the same image. We systematically explore a broad design space of visual augmentations and uncover that (1) asymmetry matters: all four augmentation families improve performance, while symmetric self-distillation degrades it; (2) strength matters: performance peaks at a moderate strength; and (3) the gap must remain task-consistent: augmentations that completely remove the question-relevant evidence can induce large but uninformative discrepancies. Across six fine-grained perception benchmarks, S^2VOPD improves Qwen3.5-4B from 70.7% to 77.4%, above all open-source models compared, up to Qwen3-VL at 235B, and surpasses GPT-5.4. While holding training data the same, it recovers 96% of the improvement achieved by methods with privileged information. Website is at https://williamium3000.github.io/s2vopd
Original Article
View Cached Full Text

Cached at: 08/17/26, 11:46 AM

Paper page - Self-Supervised Visual On-Policy Distillation

Source: https://huggingface.co/papers/2608.14144

Abstract

Self-supervised visual on-policy distillation improves small vision-language models by distilling from original images into strongly augmented student views without privileged annotations or larger teachers.

Visual on-policy distillationrelies heavily on an informativeteacher-student asymmetry, through either a larger, stronger teacher orprivileged supervision, such as reference answers or ground-truth regions of interest. This raises a fundamental question: where can informative asymmetry come from when nothing privileged is available? We answer this by inverting where the asymmetry comes from. Rather than adding privileged information to the teacher, we subtract information from the student. This asymmetry creates the same effective learning signal for free as a teacher with access to information unavailable to the student, without ground-truth annotations, rewards, or a separate stronger teacher model. Building on this principle, we introduceSelf-Supervised Visual On-Policy Distillation(S^2VOPD), a simple yet effective method that constructs on-policy learning signals fromasymmetric augmented views. S^2VOPD distills the teacher’s distribution conditioned on the original image on-policy into the student distribution conditioned on a strongly augmented view of the same image. We systematically explore a broad design space of visual augmentations and uncover that (1) asymmetry matters: all four augmentation families improve performance, while symmetricself-distillationdegrades it; (2) strength matters: performance peaks at a moderate strength; and (3) the gap must remain task-consistent: augmentations that completely remove the question-relevant evidence can induce large but uninformative discrepancies. Across sixfine-grained perceptionbenchmarks, S^2VOPD improves Qwen3.5-4B from 70.7% to 77.4%, above all open-source models compared, up to Qwen3-VL at 235B, and surpasses GPT-5.4. While holding training data the same, it recovers 96% of the improvement achieved by methods with privileged information. Website is at https://williamium3000.github.io/s2vopd

View arXiv pageView PDFProject pageAdd to collection

Get this paper in your agent:

hf papers read 2608\.14144

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2608.14144 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2608.14144 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2608.14144 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

OPD-V: Visual On-Policy Self-Distillation with Modality Balance

Hugging Face Daily Papers

Introduces OPD-V, a visual on-policy self-distillation paradigm for multimodal large language models that leverages positive and negative teachers to exploit modality balance as privileged information, improving reasoning performance across benchmarks while reducing training cost.

Weak-to-Strong On-Policy Distillation

arXiv cs.LG

Introduces Weak-to-Strong On-Policy Distillation (W2S-OPD), a framework that improves a strong language model by distilling from multiple weaker models using contrast pairs in logit space, consistently outperforming standard on-policy distillation on math and code benchmarks.

On-Policy Self-Distillation without Any Supervision

Hugging Face Daily Papers

Introduces U-OPSD, an unsupervised on-policy self-distillation method that uses internal consistency and majority-vote pseudo-solutions to improve LLMs without external supervision, matching or exceeding supervised methods on math benchmarks.

Learning to Foresee: Unveiling the Unlocking Efficiency of On-Policy Distillation

arXiv cs.CL

This paper investigates the parameter-level mechanisms behind the efficiency of On-Policy Distillation (OPD) for large language models, attributing it to early 'foresight' in module allocation and update direction. It proposes EffOPD, a plug-and-play method that accelerates OPD training by 3x without compromising final performance.