Language Modeling using LMUs: 10x Better Data Efficiency or Improved Scaling Compared to Transformers

arXiv preprint, 2021

Narsimha Chilkuri, Eric Hunsberger, Aaron Voelker, Gurshaant Malik, Chris Eliasmith

Abstract

Recent studies have demonstrated that the performance of transformers on the task of language modeling obeys a power-law relationship with model size over six orders of magnitude. While transformers exhibit impressive scaling, their performance hinges on processing large amounts of data, and their computational and memory requirements grow quadratically with sequence length. Motivated by these considerations, we construct a Legendre Memory Unit based model that introduces a general prior for sequence processing and exhibits an O(n) and O(nlnn) (or better) dependency for memory and computation respectively. Over three orders of magnitude, we show that our new architecture attains the same accuracy as transformers with 10x fewer tokens. We also show that for the same amount of training our model improves the loss over transformers about as much as transformers improve over LSTMs. Additionally, we demonstrate that adding global self-attention complements our architecture and the augmented model improves performance even further.

Full text links

 PDF

 arXiv

Preprint

Journal
arXiv preprint
Arxiv
2110.02402

Cite

Plain text

BibTeX