Augmenting Self-attention with Persistent Memory

Abstract

Transformer networks have lead to important progress in language modeling and machine translation. These models include two consecutive modules, a feed-forward layer and a self-attention layer. The latter allows the network to capture long term dependencies and are often regarded as the key ingredient in the success of Transformers. Building upon this intuition, we propose a new model that solely consists of attention layers. More precisely, we augment the self-attention layers with persistent memory vectors that play a similar role as the feed-forward layer. Thanks to these vectors, we can remove the feed-forward layer without degrading the performance of a transformer. Our evaluation shows the benefits brought by our model on standard character and word level language modeling benchmarks.

References in corpus (5)

• On the difficulty of training Recurrent Neural Networks
• Recurrent Neural Network Regularization
• Generating Long Sequences with Sparse Transformers
• Pay Less Attention with Lightweight and Dynamic Convolutions
• Improving Neural Language Models with a Continuous Cache

Cited by in corpus (5)

• Reformer: The Efficient Transformer
• MUSE: Parallel Multi-Scale Attention for Sequence to Sequence Learning
• DFSMN-SAN with Persistent Memory Model for Automatic Speech Recognition
• Is Attention All What You Need? -- An Empirical Investigation on Convolution-Based Active Memory and Self-Attention
• Adaptive Transformers in RL