ZOTTA: Test-Time Adaptation with Gradient-Free Zeroth-Order Optimization

TL;DR

ZOTTA introduces a fully gradient-free test-time adaptation framework using zeroth-order optimization, enabling efficient, architecture-agnostic model adaptation under distribution shifts without backpropagation.

Contribution

The paper proposes ZOTTA, a novel BP-free TTA method that employs distribution-robust layer selection and spatial feature alignment to improve convergence and stability in high-dimensional models.

Findings

Outperforms or matches backpropagation-based methods on multiple benchmarks.

Reduces memory usage by 84% compared to SAR.

Improves accuracy by 3.9% on ImageNet-C.

Abstract

Test-time adaptation (TTA) aims to improve model robustness under distribution shifts by adapting to unlabeled test data, but most existing methods rely on backpropagation (BP), which is computationally costly and incompatible with non-differentiable models such as quantized models, limiting practical deployment on numerous edge devices. Recent BP-free approaches alleviate overhead but remain either architecture-specific or limited in optimization capacity to handle high-dimensional models. We propose ZOTTA, a fully BP-free TTA framework that performs efficient adaptation using only forward passes via Zeroth-Order Optimization (ZOO). While ZOO is theoretically appealing, naive application leads to slow convergence under high-dimensional parameter spaces and unstable optimization due to the lack of labels. ZOTTA overcomes these challenges through 1) Distribution-Robust Layer Selection, which automatically identifies and freezes layers that already extract distribution-invariant features, updating only domain-sensitive layers to reduce the optimization dimensionality and accelerate convergence; 2) Spatial Feature Aggregation Alignment, which stabilizes ZOO by aligning globally aggregated spatial features between source and target to reduce gradient variance. Together, these components enable architecture-agnostic and stable BP-free adaptation. Extensive experiments on ImageNet-C/R/Sketch/A show that ZOTTA outperforms or matches BP-based methods, e.g., it reduces memory usage by 84% and improves accuracy by 3.9% over SAR on ImageNet-C.