Sara A. Al-Emadi, Yin Yang, Ferda Ofli

Jiansheng Li, Xingxuan Zhang, Hao Zou +6

Xingyu Wang, Tao Wang

Test-Time Adaptation (TTA) is essential for enabling deep learning models to handle real-world data distribution shifts. However, current approaches face significant limitations: backpropagation-based methods are not suitable for low-end deployment devices, due to their high computation and memory requirements, as well as their tendency to modify model weights during adaptation; while traditional backpropagation-free techniques exhibit constrained adaptation capabilities. In this work, we propose Forward-Only Zeroth-Order Optimization (FOZO), a novel and practical backpropagation-free paradigm

Yongxin Guo, Lin Wang, Xiaoying Tang +1

Nikos Efthymiadis, Giorgos Tolias, Ondřej Chum

Sameer Ambekar, Zehao Xiao, Xiantong Zhen +1