# DINOv3-Diffusion Policy: Self-Supervised Large Visual Model for Visuomotor Diffusion Policy Learning

arXiv: https://arxiv.org/abs/2509.17684 · HTML: https://mlsanity.com/paper/2509.17684/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in Diffusion and flow policies for robotics, one of 14 directions in Diffusion models.
- Builds on 20 of its 23 references that are in the map.
- First posted on arXiv 22 Sep 2025.

## Placed in

- Diffusion models / [Diffusion and flow policies for robotics](https://mlsanity.com/diffusion/diffusion-and-flow-policies-for-robotics/)

## Builds on (most cited first, 20 of 20)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Denoising Diffusion Probabilistic Models](https://mlsanity.com/paper/2006.11239/) (2020)
- [U-Net: Convolutional Networks for Biomedical Image Segmentation](https://mlsanity.com/paper/1505.04597/) (2015)
- [3D Gaussian Splatting for Real-Time Radiance Field Rendering](https://mlsanity.com/paper/2308.04079/) (2023)
- [OpenVLA: An Open-Source Vision-Language-Action Model](https://mlsanity.com/paper/2406.09246/) (2024)
- [Diffusion Policy: Visuomotor Policy Learning via Action Diffusion](https://mlsanity.com/paper/2303.04137/) (2023)
- [A Survey of Large Language Models](https://mlsanity.com/paper/2303.18223/) (2023)
- [FiLM: Visual Reasoning with a General Conditioning Layer](https://mlsanity.com/paper/1709.07871/) (2017)
- [DINOv3](https://mlsanity.com/paper/2508.10104/) (2025)
- [What Matters in Learning from Offline Human Demonstrations for Robot Manipulation](https://mlsanity.com/paper/2108.03298/) (2021)
- [ImageNet-21K Pretraining for the Masses](https://mlsanity.com/paper/2104.10972/) (2021)
- [Implicit Behavioral Cloning](https://mlsanity.com/paper/2109.00137/) (2021)
- [Universal Manipulation Interface: In-The-Wild Robot Teaching Without In-The-Wild Robots](https://mlsanity.com/paper/2402.10329/) (2024)
- [From Machine Learning to Robotics: Challenges and Opportunities for Embodied Intelligence](https://mlsanity.com/paper/2110.15245/) (2021)
- [A Survey of Embodied Learning for Object-Centric Robotic Manipulation](https://mlsanity.com/paper/2408.11537/) (2024)
- [LLM A*: Human in the Loop Large Language Models Enabled A* Search for Robotics](https://mlsanity.com/paper/2312.01797/) (2023)
- [Robot Shape and Location Retention in Video Generation Using Diffusion Models](https://mlsanity.com/paper/2407.02873/) (2024)
- [LLM Granularity for On-the-Fly Robot Control](https://mlsanity.com/paper/2406.14653/) (2024)

## Built on by (most cited first, 2 of 2)

- [Layout-Aware Representation Learning for Open-Set ID Fraud Discovery](https://mlsanity.com/paper/2605.05215/) (2026)
- [Latent evolving World Action Model](https://mlsanity.com/paper/2609.27455/) (2026)

## Abstract (arXiv)

This paper evaluates DINOv3, a recent large-scale self-supervised vision backbone, for visuomotor diffusion policy learning in robotic manipulation. We investigate whether a purely self-supervised encoder can match or surpass conventional supervised ImageNet-pretrained backbones (e.g., ResNet-18) under three regimes: training from scratch, frozen, and finetuned. Across four benchmark tasks (Push-T, Lift, Can, Square) using a unified FiLM-conditioned diffusion policy, we find that (i) finetuned DINOv3 matches or exceeds ResNet-18 on several tasks, (ii) frozen DINOv3 remains competitive, indicating strong transferable priors, and (iii) self-supervised features improve sample efficiency and robustness. These results support self-supervised large visual models as effective, generalizable perceptual front-ends for action diffusion policies, motivating further exploration of scalable label-free pretraining in robotic manipulation. Compared to using ResNet18 as a backbone, our approach with DINOv3 achieves up to a 10% absolute increase in test-time success rates on challenging tasks such as Can, and on-the-par performance in tasks like Lift, PushT, and Square.