FAAST: Forward-Only Associative Learning via Closed-Form Fast Weights for Test-Time Supervised Adaptation

2026-05-06

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.

Problem

• test-time adaptation
• supervised adaptation
• fast weights
• forward-only learning
• resource-constrained models

Innovation

• forward-only learning
• fast weights
• test-time adaptation
• closed-form solution
• memory-efficient adaptation