# Learning Actionable Manipulation Recovery via Counterfactual Failure Synthesis

arXiv: https://arxiv.org/abs/2603.13528 · HTML: https://mlsanity.com/paper/2603.13528/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- Builds on 25 of its 29 references that are in the map.
- First posted on arXiv 13 Mar 2026.

## Builds on (most cited first, 25 of 25)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Control](https://mlsanity.com/paper/2307.15818/) (2023)
- [OpenVLA: An Open-Source Vision-Language-Action Model](https://mlsanity.com/paper/2406.09246/) (2024)
- [PaLM-E: An Embodied Multimodal Language Model](https://mlsanity.com/paper/2303.03378/) (2023)
- [Genie: Generative Interactive Environments](https://mlsanity.com/paper/2402.15391/) (2024)
- [MimicGen: A Data Generation System for Scalable Robot Learning using Human Demonstrations](https://mlsanity.com/paper/2310.17596/) (2023)
- [π*0.6: a VLA That Learns From Experience](https://mlsanity.com/paper/2511.14759/) (2025)
- [RoboPoint: A Vision-Language Model for Spatial Affordance Prediction for Robotics](https://mlsanity.com/paper/2406.10721/) (2024)
- [RoboGen: Towards Unleashing Infinite Data for Automated Robot Learning via Generative Simulation](https://mlsanity.com/paper/2311.01455/) (2023)
- [RoboDreamer: Learning Compositional World Models for Robot Imagination](https://mlsanity.com/paper/2404.12377/) (2024)
- [REFLECT: Summarizing Robot Experiences for Failure Explanation and Correction](https://mlsanity.com/paper/2306.15724/) (2023)
- [Ctrl-World: A Controllable Generative World Model for Robot Manipulation](https://mlsanity.com/paper/2510.10125/) (2025)
- [SpatialBot: Precise Spatial Understanding with Vision Language Models](https://mlsanity.com/paper/2406.13642/) (2024)
- [RoboBrain: A Unified Brain Model for Robotic Manipulation from Abstract to Concrete](https://mlsanity.com/paper/2502.21257/) (2025)
- [RoboRefer: Towards Spatial Referring with Reasoning in Vision-Language Models for Robotics](https://mlsanity.com/paper/2506.04308/) (2025)
- [AHA: A Vision-Language-Model for Detecting and Reasoning Over Failures in Robotic Manipulation](https://mlsanity.com/paper/2410.00371/) (2024)
- [RoboCat: A Self-Improving Generalist Agent for Robotic Manipulation](https://mlsanity.com/paper/2306.11706/) (2023)
- [SAFE: Multitask Failure Detection for Vision-Language-Action Models](https://mlsanity.com/paper/2506.09937/) (2025)
- [RoboReward: General-Purpose Vision-Language Reward Models for Robotics](https://mlsanity.com/paper/2601.00675/) (2026)
- [SAGE: Scalable Agentic 3D Scene Generation for Embodied AI](https://mlsanity.com/paper/2602.10116/) (2026)
- [AnchorDream: Repurposing Video Diffusion for Embodiment-Aware Robot Data Synthesis](https://mlsanity.com/paper/2512.11797/) (2025)
- [Guardian: Detecting Robotic Planning and Execution Errors with Vision-Language Models](https://mlsanity.com/paper/2512.01946/) (2025)
- [ReWorld: Multi-Dimensional Reward Modeling for Embodied World Models](https://mlsanity.com/paper/2601.12428/) (2026)
- [Self-Refining Vision Language Model for Robotic Failure Detection and Reasoning](https://mlsanity.com/paper/2602.12405/) (2026)

## Built on by (most cited first, 10 of 10)

- [Back to the Familiar Future: Failure Recovery for VLA Policies via Pre-Imagined Milestone Selection](https://mlsanity.com/paper/2606.09258/) (2026)
- [Imagining Recovery: Inference-Time Counterfactual Realignment for Vision-Language-Action Models](https://mlsanity.com/paper/2608.14822/) (2026)
- [Safe Embodied AI for Long-horizon Tasks: A Cross-layer Analysis of Robotic Manipulation](https://mlsanity.com/paper/2606.05660/) (2026)
- [World Models for Embodied Intelligence: From Plausible to Controllable to Actionable](https://mlsanity.com/paper/2609.16697/) (2026)
- [In-Context Learning for Robots: Methods and Applications](https://mlsanity.com/paper/2609.36012/) (2026)
- [FailBench: How Reliable are VLMs at Judging Robot Task Success?](https://mlsanity.com/paper/2609.03611/) (2026)
- [When Should a Failing Robot Ask? Initiating Corrective Human-Robot Dialogue from Audited Sensor Evidence](https://mlsanity.com/paper/2609.21942/) (2026)
- [Anatomy of a Closed-Loop Collapse: A Causal Case Study of a Compressed VLA Policy](https://mlsanity.com/paper/2609.23048/) (2026)
- [OpenWAM: An Open Framework for Composable World-Action Models](https://mlsanity.com/paper/2610.07922/) (2026)
- [RoboAware: Learning to Coordinate Embodied Skills from Counterfactual Outcomes](https://mlsanity.com/paper/2610.11480/) (2026)

## Abstract (arXiv)

While recent foundation models have significantly advanced robotic manipulation, these systems still struggle to autonomously recover from execution errors. Current failure-learning paradigms rely on either costly and unsafe real-world data collection or simulator-based perturbations, which introduce a severe sim-to-real gap. Furthermore, existing visual analyzers predominantly output coarse, binary diagnoses rather than the executable, trajectory-level corrections required for actual recovery. To bridge the gap between failure diagnosis and actionable recovery, we introduce Dream2Fix, a framework that synthesizes photorealistic, counterfactual failure rollouts directly from successful real-world demonstrations. By perturbing actions within a generative world model, Dream2Fix creates paired failure-correction data without relying on simulators. To ensure the generated data is physically viable for robot learning, we implement a structured verification mechanism that strictly filters rollouts for task validity, visual coherence, and kinematic safety. This engine produces a high-fidelity dataset of over 120k paired samples. Using this dataset, we fine-tune a vision-language model to jointly predict failure types and precise recovery trajectories, mapping visual anomalies directly to corrective actions. Extensive real-world robotic experiments show our approach achieves state-of-the-art correction accuracy, improving from 19.7% to 81.3% over prior baselines, and successfully enables zero-shot closed-loop failure recovery in physical deployments.