# One Trajectory, Two Requirements: Accuracy and Local Stability in Euler Sampling for Flow Matching Policies

DevFeed: [One Trajectory, Two Requirements: Accuracy and Local Stability in Euler Sampling for Flow Matching Policies](<https://devfeed.tech/articles/one-trajectory-two-requirements-accuracy-and-local-stability-in-euler-sampling-for-flow-matching-policies-128689.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.13311>)

Author: Mahindra Rautela

Published: 2026-11-11T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>)

Topics: [Machine learning](<https://devfeed.tech/topics/machine-learning.md>)

Tags: [abstract](<https://devfeed.tech/tags/abstract.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [can](<https://devfeed.tech/tags/can.md>), [cases](<https://devfeed.tech/tags/cases.md>), [checkpoint](<https://devfeed.tech/tags/checkpoint.md>), [condition](<https://devfeed.tech/tags/condition.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [cs-ro](<https://devfeed.tech/tags/cs-ro.md>), [error](<https://devfeed.tech/tags/error.md>), [exhibits](<https://devfeed.tech/tags/exhibits.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [fixed](<https://devfeed.tech/tags/fixed.md>), [flow](<https://devfeed.tech/tags/flow.md>), [flow-matching](<https://devfeed.tech/tags/flow-matching.md>), [inference](<https://devfeed.tech/tags/inference.md>), [input](<https://devfeed.tech/tags/input.md>), [local](<https://devfeed.tech/tags/local.md>), [matching](<https://devfeed.tech/tags/matching.md>), [one](<https://devfeed.tech/tags/one.md>), [policies](<https://devfeed.tech/tags/policies.md>), [requirements](<https://devfeed.tech/tags/requirements.md>), [sampling](<https://devfeed.tech/tags/sampling.md>)

## AI overview

A step count that tracks a flow matching policy's nominal noise-to-action trajectory may still be too coarse for local Euler stability: controlled experiments with an exact two-coordinate system and six trained flow models show that accuracy and stability can require different step counts without changing the learned transport. In a frozen π₀.₅-LIBERO checkpoint, 30 inputs across 10 tasks showed strongly contracting endpoint directions, but no sampled mode violated the study's Euler non-expansion condition. For five inputs, accuracy sweeps needed 4-9 steps for 5% terminal error and 87-112 for 1%; these cases are consistent with accuracy-limited sampling under that local criterion. The findings caution that endpoint accuracy alone does not establish whether stability imposes a separate requirement.

## Source excerpt

arXiv:2611.13311v1 Announce Type: new Abstract: Flow matching robot policies commonly use a fixed Euler step count selected from offline accuracy sweeps. At inference time, however, each policy input (visual observation, robot state, and language instruction) together with a source-noise draw defines a different ODE trajectory and sequence of local Jacobians. We ask whether a step count that accurately follows the nominal noise-to-action trajectory can still be too coarse for strongly contracting directions around it. An exact two-coordinate system and a controlled augmentation of six trained flow models show that trajectory accuracy and local Euler stability can impose different step requirements even when the original learned transport is unchanged. We then study a frozen $\pi_{0.5}$-LIBERO checkpoint without modifying its weights or action dynamics. Across 30 policy inputs from 10 LIBERO-10 tasks, the Jacobian exhibits strong contracting directions near the action endpoint, yet no sampled mode violates the Euler non-expansion condition. On five of those inputs, accuracy sweeps require 4--9 Euler steps at 5% terminal error and 87--112 at 1%. These five paired cases are therefore consistent with accuracy-limited rather than stability-limited sampling under this local criterion, while the controlled experiments show why endpoint accuracy alone cannot rule out a separate stability requirement.