# SI-Diff: A Framework for Learning Search and High-Precision Insertion with a Force-Domain Diffusion Policy

arXiv: https://arxiv.org/abs/2605.12247 · HTML: https://mlsanity.com/paper/2605.12247/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Diffusion and flow policies for robotics, one of 14 directions in Diffusion models.
- Builds on 22 of its 35 references that are in the map.
- First posted on arXiv 12 May 2026.

## 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, 22 of 22)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Imitating Human Behaviour with Diffusion Models](https://mlsanity.com/paper/2301.10677/) (2023)
- [Balanced Contrastive Learning for Long-Tailed Visual Recognition](https://mlsanity.com/paper/2207.09052/) (2022)
- [Deep Reinforcement Learning for High Precision Assembly Tasks](https://mlsanity.com/paper/1708.04033/) (2017)
- [FORGE: Force-Guided Exploration for Robust Contact-Rich Manipulation under Uncertainty](https://mlsanity.com/paper/2408.04587/) (2024)
- [TacDiffusion: Force-domain Diffusion Policy for Precise Tactile Manipulation](https://mlsanity.com/paper/2409.11047/) (2024)
- [Learning Insertion Primitives with Discrete-Continuous Hybrid Action Space for Robotic Assembly Tasks](https://mlsanity.com/paper/2110.12618/) (2021)
- [Active Extrinsic Contact Sensing: Application to General Peg-in-Hole Insertion](https://mlsanity.com/paper/2110.03555/) (2021)
- [Vision-driven Compliant Manipulation for Reliable, High-Precision Assembly Tasks](https://mlsanity.com/paper/2106.14070/) (2021)
- [MV-DeepSDF: Implicit Modeling with Multi-Sweep Point Clouds for 3D Vehicle Reconstruction in Autonomous Driving](https://mlsanity.com/paper/2309.16715/) (2023)
- [UniGaussian: Driving Scene Reconstruction from Multiple Camera Models via Unified Gaussian Representations](https://mlsanity.com/paper/2411.15355/) (2024)
- [Fast robust peg-in-hole insertion with continuous visual servoing](https://mlsanity.com/paper/2011.06399/) (2020)
- [VQA-Diff: Exploiting VQA and Diffusion for Zero-Shot Image-to-3D Vehicle Asset Generation in Autonomous Driving](https://mlsanity.com/paper/2407.06516/) (2024)
- [Application of Ghost-DeblurGAN to Fiducial Marker Detection](https://mlsanity.com/paper/2109.03379/) (2021)
- [Intensity Image-based LiDAR Fiducial Marker System](https://mlsanity.com/paper/2203.01816/) (2022)
- [HIPPo: Harnessing Image-to-3D Priors for Model-free Zero-shot 6D Pose Estimation](https://mlsanity.com/paper/2502.10606/) (2025)
- [Autonomous Vision-based UAV Landing with Collision Avoidance using Deep Learning](https://mlsanity.com/paper/2109.08628/) (2021)
- [POMDP-Guided Active Force-Based Search for Robotic Insertion](https://mlsanity.com/paper/2404.03943/) (2024)
- [L-PR: Exploiting LiDAR Fiducial Marker for Unordered Low Overlap Multiview Point Cloud Registration](https://mlsanity.com/paper/2406.03298/) (2024)
- [Perception-Control Coupled Visual Servoing for Textureless Objects Using Keypoint-Based EKF](https://mlsanity.com/paper/2602.06834/) (2026)

## Abstract (arXiv)

Contact-rich assembly is fundamental in robotics but poses significant challenges due to uncertainties in relative poses, such as misalignments and small clearances in peg-in-hole tasks. Existing approaches typically address search and high-precision insertion separately, because these tasks involve distinct action patterns. However, supporting both tasks within a single model, without switching models or weights, is desirable for intelligent assembly systems. In this work, we propose SI-Diff, a framework that learns both search and high-precision insertion through a force-domain diffusion policy. To this end, we introduce a new mode-conditioning mechanism that enables the policy to capture distinct action behaviors under a single framework. Moreover, we develop a new search teacher policy that can generate diverse trajectories. By training on successful and efficient demonstrations provided by the teacher policy, the model learns the mapping from tactile and end-effector velocity observations to effective action behaviors. We conduct thorough experiments to show that SI-Diff extends the tolerance to x-y misalignments from 2 mm to 5 mm compared to the state-of-the-art baseline, TacDiffusion, while also demonstrating strong zero-shot transferability to unseen shapes.