一句话：提出路由熵，利用 MoE 模型内部路由分布的不确定性检测分布外数据，并可与现有输出方法互补。
与训练数据差异显著的未见数据会使机器学习模型行为不可预测，在自动驾驶等安全关键系统中尤为危险，因此分布外（OOD）检测至关重要。现有方法多依赖模型最终输出，而模型可能对未见数据过度自信，导致检测受限。本文提出路由熵（Routing Entropy）：一种利用专家混合（MoE）模型内部路由行为的新型 OOD 检测方法。其假设是，MoE 对分布内输入会给出高置信度的路由，而对 OOD 输入则表现出更大的不确定性；作者通过对给定输入的路由分数计算熵来量化这种不确定性。在用于自动驾驶感知的基于 MoE 的语义分割模型上实验表明，路由熵单独使用即有效，更重要的是它能与现有基于输出的方法提供互补信号：将路由熵与现有方法结合可显著提升 OOD 检测性能。这表明利用 MoE 内部路由行为是通向鲁棒 OOD 检测的有前景方向。
Unseen data that differ significantly from the training data can cause machine learning models to behave unpredictably, which is particularly problematic in safety-critical systems like autonomous vehicles. Detecting such data, commonly called out-of-distribution (OOD) data, is essential for ensuring the robustness of these models. Existing methods often rely on the model’s final output, which are limited since the model can be overconfident on unseen data. In this paper, we propose Routing Entropy, a novel OOD detection method that leverages the internal routing behavior of Mixture-of-Experts (MoE) models, a design increasingly adopted in modern neural networks. We hypothesize that MoE models exhibit high confidence routing for in-distribution (ID) inputs, but greater uncertainty for OOD inputs. We quantify this uncertainty by calculating the entropy of the routing scores for a given input. Experimental results on a MoE-based semantic segmentation model used for perception in autonomous driving demonstrate that Routing Entropy is effective on its own and, more importantly, provides a complementary signal to existing output-based methods. Combining Routing Entropy with an existing method significantly improves OOD detection performance. These results suggest that leveraging internal routing behavior of MoE models is a promising direction for robust OOD detection.
DOI 原文 ·
@inproceedings{paperbot9213,
title = {Unseen Data Detection using Routing Entropy in Mixture-of-Experts for Autonomous Vehicles},
author = {Sang In Lee and Donghwan Shin and Jihun Park},
booktitle = {ASE},
year = {2025},
doi = {10.1109/ase63991.2025.00332}
}