mlsanityData to 14 Oct 2026

MemCorr-DP: Counterfactual Correspondence Conditioning for a Diffusion Policy Guided by a Reference

arXiv 2609.06615 ↗First posted 6 Sep 2026cs.RO cs.AIcited by 0

On the map: Diffusion models → Diffusion and flow policies for robotics

Abstract (arXiv)

Behavior-cloned visuomotor policies can remain accurate near their training distribution yet fail when object position and camera viewpoint change together. A successful reference trajectory contains the geometry needed to transfer the same interaction, but the policy must align that geometry with the current scene and remain sensitive to it during denoising. To address these challenges, we present MemCorr-DP, a diffusion policy that lifts frozen RoMa v2 matches into explicit 3D relations between the current scene and the reference trajectory. A counterfactual paired objective assigns opposite behaviors the same physical state and noisy action while retaining reference-specific denoising targets. Mixed-condition fine-tuning then adapts the policy from ground-truth geometry to measured correspondence errors. Our strongest evaluation places the Door in the outermost position bands beyond the training support and changes the query camera by ±15^circ. Under this combined shift, MemCorr-DP achieves 96.67% closed-loop success, compared with 88.00% for a visual Transformer with the same action architecture. Objective ablations and reference interventions show that behavior responds to the selected reference, while matched controls favor the complete relation set over future motion or centroid geometry alone. These results support explicit 3D reference relations as a robust conditioning interface when spatial and viewpoint changes are compounded in the evaluated task.

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 24 of its 32 references that are in the map. First posted on arXiv 6 Sep 2026.

Builds on

24 of 32 references.

Closest ancestors · 8

Same direction, newest first

  1. 2025Bridging Perception and Action: Spatially-Grounded Mid-Level Representations for Robot Generalization6 Jun
  2. 2025A Practical Guide for Incorporating Symmetry in Diffusion Policy19 May
  3. 2025Conditioning Matters: Training Diffusion Policies is Faster Than You Think16 May
  4. 2024AffordDP: Generalizable Diffusion Policy with Transferable Affordance4 Dec
  5. 2024GenDP: 3D Semantic Fields for Category-Level Generalizable Diffusion Policy23 Oct
  6. 2024EquiBot: SIM(3)-Equivariant Diffusion Policy for Generalizable and Data Efficient Learning1 Jul
  7. 20243D Diffuser Actor: Policy Diffusion with 3D Scene Representations16 Feb
  8. 2023Diffusion Policy: Visuomotor Policy Learning via Action Diffusion7 Mar

Other references · 16

  1. 2017Attention Is All You Needcited by 60,361
  2. 2020Denoising Diffusion Probabilistic Modelscited by 21,076
  3. 2020Denoising Diffusion Implicit Modelscited by 9,781
  4. 2025DINOv3cited by 1,414
  5. 2019Meta-World: A Benchmark and Evaluation for Multi-Task and Meta Reinforcement Learningcited by 1,204
  6. 2023RoMa: Robust Dense Feature Matchingcited by 302
  7. 2022DKM: Dense Kernelized Feature Matching for Geometry Estimationcited by 177
  8. 2024Flow as the Cross-Domain Manipulation Interfacecited by 132
  9. 2025RoMa v2: Harder Better Faster Denser Feature Matchingcited by 51
  10. 2022TAX-Pose: Task-Specific Cross-Pose Estimation for Robot Manipulationcited by 47
  11. 2024Render and Diffuse: Aligning Image and Action Spaces for Diffusion-based Behaviour Cloningcited by 20
  12. 2025CordViP: Correspondence-based Visuomotor Policy for Dexterous Manipulation in Real-Worldcited by 18
4 more
  1. 2023One-shot Imitation Learning via Interaction Warpingcited by 16
  2. 2025Knowledge-Driven Imitation Learning: Enabling Generalization Across Diverse Conditionscited by 2
  3. 2025Robust Learning of Diffusion Models with Extremely Noisy Conditionscited by 2
  4. 2026MatchingPolicy: Correspondence-Aware Policy Enables Cross-Object In-Context Learningcited by 2

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Hyper-DP3: Frequency-Aware Right-Sizing of 3D Diffusion Policies for Visuomotor Control
  2. 2026Spatially Conditioned Diffusion Policy: Learning Precise and Robust Manipulation with a Single RGB Camera
  3. 2025Freeze, Share, Shrink: Rethinking the Action Backbone in Diffusion Policies
  4. 2026Immiscible Diffusion Policy: Preserving Multimodal Robot Actions through Label-Free Noise Assignment
  5. 2025ADPro: a Test-time Adaptive Diffusion Policy via Manifold-constrained Denoising and Task-aware Initialization for Robotic Manipulation
  6. 2025D3P: Dynamic Denoising Diffusion Policy via Reinforcement Learning
  7. 2025ISS Policy : Scalable Diffusion Policy with Implicit Scene Supervision
  8. 20253D Flow Diffusion Policy: Visuomotor Policy Learning via Generating Flow in 3D Space