mlsanityData to 14 Oct 2026

MAGMA-GEN: Validated Recovery Supervision from Ambiguous Failures via Counterfactual Re-Execution

arXiv 2609.20056 ↗First posted 17 Sep 2026cs.AIcited by 2

Not on the map. Why?

Abstract (arXiv)

Hierarchical robotic systems executing long-horizon manipulation tasks must make high-level semantic decisions that orchestrate stochastic low-level skills. In this setting, failed rollouts are ambiguous: a poor downstream state may reflect an invalid high-level decision, partial observation, or a valid decision whose physical execution failed. Traditional supervised learning lacks data for such recovery states, while reinforcement learning struggles with sparse rewards and non-local credit assignment. We propose MAGMA-GEN, an on-policy data-generation pipeline that converts ambiguous failed rollouts into validated recovery supervision. MAGMA-GEN first uses a privileged coach to hypothesize an early decision-level error and propose localized correction or recovery actions. Because this diagnosis is fallible, candidates are retained only if re-execution from the same state under matched conditions improves downstream progress. This produces supervised examples from the agent's own failure distribution without per-step human demonstrations. Evaluated on interactive long-horizon manipulation tasks, MAGMA-GEN improves task success and recovery capabilities, against distillation and trajectory-repair baselines under evolving task constraints in both simulation and real-robot execution.

Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Builds on 32 of its 40 references that are in the map. First posted on arXiv 17 Sep 2026.

Builds on

32 of 40 references.

References · 32

  1. 2024π0: A Vision-Language-Action Flow Model for General Robot Controlcited by 2,407
  2. 2022Do As I Can, Not As I Say: Grounding Language in Robotic Affordancescited by 2,348
  3. 2023PaLM-E: An Embodied Multimodal Language Modelcited by 2,071
  4. 2023Learning Fine-Grained Bimanual Manipulation with Low-Cost Hardwarecited by 1,568
  5. 2022Code as Policies: Language Model Programs for Embodied Controlcited by 1,210
  6. 2025Does Reinforcement Learning Really Incentivize Reasoning Capacity in LLMs Beyond the Base Model?cited by 982
  7. 2019ALFRED: A Benchmark for Interpreting Grounded Instructions for Everyday Taskscited by 783
  8. 2022ProgPrompt: Generating Situated Robot Task Plans using Large Language Modelscited by 611
  9. 2021CALVIN: A Benchmark for Language-Conditioned Policy Learning for Long-Horizon Robot Manipulation Taskscited by 584
  10. 2022LLM-Planner: Few-Shot Grounded Planning for Embodied Agents with Large Language Modelscited by 462
  11. 2025LLMs Get Lost In Multi-Turn Conversationcited by 429
  12. 2025ReTool: Reinforcement Learning for Strategic Tool Use in LLMscited by 392
20 more
  1. 2023ToRA: A Tool-Integrated Reasoning Agent for Mathematical Problem Solvingcited by 282
  2. 2024ManiSkill3: GPU Parallelized Robotics Simulation and Rendering for Generalizable Embodied AIcited by 229
  3. 2025Spurious Rewards: Rethinking Training Signals in RLVRcited by 209
  4. 2023ART: Automatic multi-step reasoning and tool-use for large language modelscited by 197
  5. 2024Trial and Error: Exploration-Based Trajectory Optimization for LLM Agentscited by 197
  6. 2024VLABench: A Large-Scale Benchmark for Language-Conditioned Robotics Manipulation with Long-Horizon Reasoning Taskscited by 175
  7. 2024PARTNR: A Benchmark for Planning and Reasoning in Embodied Multi-agent Taskscited by 82
  8. 2025Agent0: Unleashing Self-Evolving Agents from Zero Data via Tool-Integrated Reasoningcited by 75
  9. 2024Real-Time Anomaly Detection and Reactive Planning with Large Language Modelscited by 68
  10. 2026Hindsight Credit Assignment for Long-Horizon LLM Agentscited by 65
  11. 2024Plan-Seq-Learn: Language Model Guided RL for Solving Long Horizon Robotics Taskscited by 52
  12. 2023M2T2: Multi-Task Masked Transformer for Object-centric Pick and Placecited by 41
  13. 2025RaC: Robot Learning for Long-Horizon Tasks by Scaling Recovery and Correctioncited by 39
  14. 2024Towards Generalizable Vision-Language Robotic Manipulation: A Benchmark and LLM-guided 3D Policycited by 38
  15. 2025Guided Self-Evolving LLMs with Minimal Human Supervisioncited by 35
  16. 2025Advancing Tool-Augmented Large Language Models via Meta-Verification and Reflection Learningcited by 24
  17. 2025Fail2Progress: Learning from Real-World Robot Failures with Stein Variational Inferencecited by 8
  18. 2025Pinpointing crucial steps: Attribution-based Credit Assignment for Verifiable Reinforcement Learningcited by 4
  19. 2025Hierarchical DLO Routing with Reinforcement Learning and In-Context Vision-language Modelscited by 2
  20. 2026Guided by Trajectories: Repairing and Rewarding Tool-Use Trajectories for Tool-Integrated Reasoningcited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026MAGMA-Bench: Evaluating Long-Horizon Embodied Agents under Dynamic Tasks and Execution Failurescited by 0
  2. 2026Flow-Support-Weighted Proposals for Test-Time Verified VLA Policiescited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026Kintsugi-VLA: Turning Failed Robot Rollouts into Recovery Data through Interventional Recoverability
  2. 2024RecoveryChaining: Learning Local Recovery Policies for Robust Manipulation
  3. 2026FAR: Failure-Aware Retry for Test-Time Recovery and Continual Policy Improvement
  4. 2026RESETTLE: Robotic Recovery through Disagreement-Triggered Retrieval and Efficient Corrective Control ↗
  5. 2026CARE: Experience-Guided Atomic Corrective Execution for Vision-Language-Action Policies
  6. 2026Learning Actionable Manipulation Recovery via Counterfactual Failure Synthesis
  7. 2026From Reaction to Anticipation: Proactive Failure Recovery through Agentic Task Graph for Robotic Manipulation
  8. 2026Learning Robust Execution in Robotic Manipulation with Agentic Reinforcement Learning