FAR: Failure-Aware Retry for Test-Time Recovery and Continual Policy Improvement
CoRL 2026
Carnegie Mellon University
Abstract
Robot policies inevitably encounter failures when deployed in real environments. Naive retries often repeat
the same mistakes, while many existing recovery methods rely on human intervention. In this paper, we propose
Failure-Aware Retry (FAR), a framework that enables robots to learn from previous
failures at test time, adapt their behavior accordingly, and eventually complete the task autonomously. FAR
combines Failure-Contrastive Preference Adaptation, which constructs preference learning data from failures to
steer the policy away from previously unsuccessful behaviors, with lightweight action perturbations during
retries to encourage local exploration. We further incorporate successful recovery trajectories into a
training loop for continual policy improvement. Experiments in both simulation and real-world manipulation
tasks show that FAR substantially improves success rates and robustness, with average gains of 17.6% over the
standard diffusion policy in simulation and 11.7% in the real world. In addition, FAR significantly improves
data efficiency under both reset and timestep budgets during continual policy improvement by exploiting
informative failure cases.
Contributions
- We introduce Failure-Aware Retry (FAR), a retry-based test-time recovery framework that combines failure-aware adaptation with lightweight exploration for more effective failure recovery.
- We show that FAR generates informative recovery trajectories that provide useful supervision for continual policy improvement.
- We demonstrate that FAR improves task success rates in both simulation and the real world, while reducing the need for costly environment resets during online learning.
Method
After a failure, FAR identifies failure-inducing actions using value estimation, then updates the policy with both failure examples and alternative positive examples. The collected trajectories are added to the replay buffer for continual policy improvement.
Evaluation & Results
Real-World
Drawer: Push the handle and close the drawer
DP-NR
DP-FAR (Ours)
Tea: Pick up the teapot and pour water into the cup
DP-NR
DP-FAR (Ours)
Pot: Pick up the lid and place it on the pot
DP-BGR
DP-FAR (Ours)
Key Results
- Improved the average success rate by 17.6% over the standard diffusion policy in simulation.
- Improved success rate by 11.7% on real-world manipulation tasks.
- FAR significantly improves data efficiency under both reset and timestep budgets during continual policy improvement by exploiting informative failure cases.
Citation
@misc{hao2026far,
title={FAR: Failure-Aware Retry for Test-Time Recovery and Continual Policy Improvement},
author={Haoran Hao and Shahram Najam Syed and Jeffrey Ichnowski and Jeff Schneider},
year={2026},
eprint={2607.01111},
archivePrefix={arXiv},
primaryClass={cs.RO},
url={https://arxiv.org/abs/2607.01111},
}