- Rating
- 2
- Confidence
- 3
- Informativeness
- 2
- Presentation
- 2
- Soundness
- 2
Read review
Paper summary
This paper introduces PICES, a training-free inference method for flow-matching VLA policies that generates multiple candidate action chunks and their perturbed counterparts in a single batched policy inference. It measures action sensitivity to perturbations and uses the scores to jointly select the action chunk and the execution horizon before replanning. The method exploits the memory-bound nature of action experts on Jetson hardware, where the lowered total inference time by reducing the number of policy calls outweighs the relatively little overhead incurred by generating additional candidates. Experiments on simulations and a real-world robot report fewer policy calls with comparable or slightly improved success rates relative to fixed-horizon execution, achieving inference-time speedups.
Critical strengths
- The paper backs up its main idea: spending a few extra samples per call to make fewer calls. It shows that choosing well among multiple sampled chunks allows longer execution. It measures on the Jetson hardware that the extra samples add only a small delay, a cost already included in its reported speedups.
- On SimplerEnv and on the real robot, PICES matches the success rate of the best fixed horizon with fewer policy calls. The ablation shows that the sensitivity-based horizon rule works very well. With a single candidate, it already accounts for most of the reduction in calls.
Critical weaknesses
- PICES does not remove the need to tune the execution horizon. On LIBERO with , a fixed horizon of 8 matches or outperforms the default PICES. On LIBERO and on the real robot, PICES outperforms the best fixed horizon in both success and speed only when its minimum horizon is set to 8, which is the best fixed horizon in the same evaluations.
- The success gains over fixed horizons are small, such as 68.7 against 68.5 on SimplerEnv and 93.5 against 90.7 on the real robot, and no error bars or repeated runs are reported. The claim of improving the success rate should read as matching it.
Questions for authors
- On the real robot, is the best fixed horizon from the sweep, and is also the shortest horizon PICES can choose. What does PICES reach with a smaller , so that the rule has to find the right horizon on its own?
- Table 1 averages policy calls over successful episodes only, so methods with different success rates are averaged over different episodes. How do call counts and speedups change when including failed episodes as well?