Search papers, labs, and topics across Lattice.
This paper introduces the Reliable Neural Collapse (ReNC) approximation for Open-World Test-Time Adaptation (OWTTA), addressing the limitations of traditional TTA methods in scenarios with label distribution shifts. By utilizing neural collapse as a structural prior, the authors show that pre-trained classifier weights can effectively serve as source domain prototypes, enabling the filtering of Out-Of-Distribution (OOD) samples for more reliable updates. Extensive experiments reveal that ReNC not only improves adaptation performance in open-world settings but also preserves essential neural collapse properties, offering valuable insights for future model design.
Pre-trained classifier weights can act as robust prototypes for adapting to shifting label distributions in open-world scenarios, significantly enhancing model reliability.
Test-Time Adaptation (TTA) methods aim to bridge the domain gap between the source and target domains. However, traditional TTA methods become ineffective when the label distribution shift occurs, a challenge commonly referred to as an open-world scenario. In this paper, we introduce a new method named Reliable Neural Collapse approximation (ReNC) for Open-World Test-Time Adaptation (OWTTA). Specifically, we leverage neural collapse as a structural prior for reliable target-domain adaptation. Guided by this prior, we justify that the pre-trained classifier weights can serve as the prototypes of the source domain. By measuring the similarity between samples and prototypes, we filter out the Out-Of-Distribution~(OOD) samples for reliable updates. Furthermore, we propose a neural collapse approximation mechanism to refine these prototypes, ensuring they can gradually adapt to the target domain while maintaining the neural collapse structure. Extensive experiments on several open-world benchmarks demonstrate the superiority of the proposed method. Our empirical analysis suggests that ReNC better preserves NC-related properties in the target domain, providing useful evidence for explaining reliable OWTTA and offering new insights for model design. Code is available at https://github.com/JiaqiLin-AI/ReNC.
