# Widening the Representation Bottleneck in Neural Machine Translation with Lexical Shortcuts

## TL;DR

This paper introduces lexical shortcut connections in Transformer models for neural machine translation, enabling dynamic access to lexical features and improving translation quality across multiple language pairs.

## Contribution

It proposes a novel gating mechanism with shortcut connections that alleviates the lexical representation bottleneck in Transformers, enhancing translation performance.

## Key findings

- Achieves an average of 0.9 BLEU improvement on WMT translation tasks.
- Reduces lexical information passing through hidden layers.
- Demonstrates the effectiveness of lexical shortcuts through ablation studies.

## Abstract

The transformer is a state-of-the-art neural translation model that uses attention to iteratively refine lexical representations with information drawn from the surrounding context. Lexical features are fed into the first layer and propagated through a deep network of hidden layers. We argue that the need to represent and propagate lexical features in each layer limits the model's capacity for learning and representing other information relevant to the task. To alleviate this bottleneck, we introduce gated shortcut connections between the embedding layer and each subsequent layer within the encoder and decoder. This enables the model to access relevant lexical content dynamically, without expending limited resources on storing it within intermediate states. We show that the proposed modification yields consistent improvements over a baseline transformer on standard WMT translation tasks in 5 translation directions (0.9 BLEU on average) and reduces the amount of lexical information passed along the hidden layers. We furthermore evaluate different ways to integrate lexical connections into the transformer architecture and present ablation experiments exploring the effect of proposed shortcuts on model behavior.