Attention Approximates Sparse Distributed Memory

Trenton Bricken; Cengiz Pehlevan

NeurIPS 2021poster44 citations

Attention Approximates Sparse Distributed Memory

Abstract

While Attention has come to be an important mechanism in deep learning, there remains limited intuition for why it works so well. Here, we show that Transformer Attention can be closely related under certain data conditions to Kanerva's Sparse Distributed Memory (SDM), a biologically plausible associative memory model. We confirm that these conditions are satisfied in pre-trained GPT2 Transformer models. We discuss the implications of the Attention-SDM map and provide new computational and biological interpretations of Attention.

TransformerAttentionSparse Distributed MemoryAssociative MemoryCerebellum

BibTeX

@inproceedings{
bricken2021attention,
title={Attention Approximates Sparse Distributed Memory},
author={Trenton Bricken and Cengiz Pehlevan},
booktitle={Advances in Neural Information Processing Systems},
editor={A. Beygelzimer and Y. Dauphin and P. Liang and J. Wortman Vaughan},
year={2021},
url={https://openreview.net/forum?id=WVYzd7GvaOM}
}