| title | FAAST: Forward-Only Associative Learning via Closed-Form Fast Weights for Test-Time Supervised Adaptation |
|---|---|
| authors | Guangsheng Bao, Hongbo Zhang, Han Cui, Ke Sun, Yanbin Zhao, Juncai He, Yue Zhang |
| date | 2026-05-08 |
| venue | arXiv (preprint, not peer reviewed) |
| id | arXiv:2605.04651 |
| affiliation | Text Intelligence Lab of Westlake University |
| upvotes | 1 (Hugging Face Daily Papers, 2026-11-01) |
| links | abs pdf hf code |
| abstract | Adapting pretrained models typically involves a trade-off between the high training costs of backpropagation and the heavy inference overhead of memory-based or in-context learning. We propose FAAST, a forward-only associative adaptation method that analytically compiles labeled examples into fast weights in a single pass. By eliminating memory or context dependence, FAAST achieves constant-time inference and decouples task adaptation from pretrained representation. Across image classification and language modeling benchmarks, FAAST matches or exceeds backprop-based adaptation while reducing adaptation time by over 90% and is competitive to memory/context-based adaptation while saving memory usage by up to 95%. These results demonstrate FAAST as a highly efficient, scalable solution for supervised task adaptation, particularly for resource-constrained models. We release the code and models at https://github.com/baoguangsheng/faast. |
cite
@misc{bao2026faast,
title={FAAST: Forward-Only Associative Learning via Closed-Form Fast Weights for Test-Time Supervised Adaptation},
author={Guangsheng Bao and Hongbo Zhang and Han Cui and Ke Sun and Yanbin Zhao and Juncai He and Yue Zhang},
year={2026},
eprint={2605.04651},
archivePrefix={arXiv},
url={https://arxiv.org/abs/2605.04651}
}