On the Robustness of Open-World Test-Time Training: Self-Training with Dynamic Prototype Expansion

TL;DR

This paper enhances open-world test-time training robustness by developing adaptive OOD pruning and dynamic prototype expansion, significantly improving performance on benchmarks with out-of-distribution data.

Contribution

It introduces a novel approach combining adaptive strong OOD pruning and dynamic prototype expansion to better distinguish and handle out-of-distribution samples during test-time training.

Findings

Achieved state-of-the-art results on 5 OWTTT benchmarks.

Improved separation of weak and strong OOD samples.

Enhanced robustness of test-time training under contaminated target domains.

Abstract

Generalizing deep learning models to unknown target domain distribution with low latency has motivated research into test-time training/adaptation (TTT/TTA). Existing approaches often focus on improving test-time training performance under well-curated target domain data. As figured out in this work, many state-of-the-art methods fail to maintain the performance when the target domain is contaminated with strong out-of-distribution (OOD) data, a.k.a. open-world test-time training (OWTTT). The failure is mainly due to the inability to distinguish strong OOD samples from regular weak OOD samples. To improve the robustness of OWTTT we first develop an adaptive strong OOD pruning which improves the efficacy of the self-training TTT method. We further propose a way to dynamically expand the prototypes to represent strong OOD samples for an improved weak/strong OOD data separation. Finally, we regularize self-training with distribution alignment and the combination yields the state-of-the-art performance on 5 OWTTT benchmarks. The code is available at https://github.com/Yushu-Li/OWTTT.