Hierarchical Adaptive networks with Task vectors for Test-Time Adaptation

TL;DR

This paper introduces Hi-Vec, a hierarchical adaptive network with task vectors that enhances test-time adaptation by dynamically selecting and merging layers, improving robustness across diverse distribution shifts.

Contribution

The paper presents a novel hierarchical adaptation method with dynamic layer selection, weight merging, and a gating mechanism to improve test-time adaptation performance.

Findings

Hi-Vec outperforms existing methods on multiple datasets.

It improves robustness against distribution shifts.

It effectively handles limited batch sizes and outliers.

Abstract

Test-time adaptation allows pretrained models to adjust to incoming data streams, addressing distribution shifts between source and target domains. However, standard methods rely on single-dimensional linear classification layers, which often fail to handle diverse and complex shifts. We propose Hierarchical Adaptive Networks with Task Vectors (Hi-Vec), which leverages multiple layers of increasing size for dynamic test-time adaptation. By decomposing the encoder's representation space into such hierarchically organized layers, Hi-Vec, in a plug-and-play manner, allows existing methods to adapt to shifts of varying complexity. Our contributions are threefold: First, we propose dynamic layer selection for automatic identification of the optimal layer for adaptation to each test batch. Second, we propose a mechanism that merges weights from the dynamic layer to other layers, ensuring all layers receive target information. Third, we propose linear layer agreement that acts as a gating function, preventing erroneous fine-tuning by adaptation on noisy batches. We rigorously evaluate the performance of Hi-Vec in challenging scenarios and on multiple target datasets, proving its strong capability to advance state-of-the-art methods. Our results show that Hi-Vec improves robustness, addresses uncertainty, and handles limited batch sizes and increased outlier rates.