Low-Rank Bottleneck in Multi-head Attention Models
This addresses the problem of large model sizes in Transformers for NLP, making them more efficient for downstream tasks, though it is an incremental improvement.
The paper identifies a low-rank bottleneck in multi-head attention models caused by the scaling between the number of heads and head size, which limits model efficiency. As a solution, it proposes setting head size to input sequence length, enabling training with smaller embedding dimensions and better performance scaling.
Attention based Transformer architecture has enabled significant advances in the field of natural language processing. In addition to new pre-training techniques, recent improvements crucially rely on working with a relatively larger embedding dimension for tokens. Unfortunately, this leads to models that are prohibitively large to be employed in the downstream tasks. In this paper we identify one of the important factors contributing to the large embedding size requirement. In particular, our analysis highlights that the scaling between the number of heads and the size of each head in the current architecture gives rise to a low-rank bottleneck in attention heads, causing this limitation. We further validate this in our experiments. As a solution we propose to set the head size of an attention unit to input sequence length, and independent of the number of heads, resulting in multi-head attention layers with provably more expressive power. We empirically show that this allows us to train models with a relatively smaller embedding dimension and with better performance scaling.