Predicting Out-of-Distribution Error with Confidence Optimal Transport

Yuzhe Lu, Zhenlin Wang, Runtian Zhai, Soheil Kolouri, Joseph Campbell, Katia P. Sycara

Keywords: Distribution Shift, OOD Error Prediction, Optimal Transport, Deep Learning

TL;DR: We present a simple yet effective method based on optimal transport to predict a model's performance on an unknown distribution without any additional annotation.

Abstract: Out-of-distribution (OOD) data poses serious challenges in deployed machine learning models as even subtle changes could incur significant performance drops. Being able to estimate a model's performance on test data is important in practice as it indicates when to trust a model's decisions. We present a simple yet effective method to predict a model's performance on an unknown distribution without any additional annotation. Our approach is rooted in the Optimal Transport theory, viewing test samples' output softmax scores from deep neural networks as empirical samples from an unknown distribution. We show that our method, Confidence Optimal Transport (COT), provides robust estimates of a model's performance on a target domain. Despite its simplicity, our method achieves state-of-the-art results on three benchmark datasets and outperforms existing methods by a large margin.