# Training Neural Machine Translation To Apply Terminology Constraints

## TL;DR

This paper introduces a training-based approach for neural machine translation that effectively incorporates custom terminology at runtime without additional computational overhead, outperforming constrained decoding methods.

## Contribution

The paper presents a novel training method enabling neural MT systems to naturally incorporate terminology constraints, reducing computational costs and increasing robustness.

## Key findings

- Outperforms state-of-the-art constrained decoding methods
- Maintains speed comparable to unconstrained decoding
- More effective in realistic translation scenarios

## Abstract

This paper proposes a novel method to inject custom terminology into neural machine translation at run time. Previous works have mainly proposed modifications to the decoding algorithm in order to constrain the output to include run-time-provided target terms. While being effective, these constrained decoding methods add, however, significant computational overhead to the inference step, and, as we show in this paper, can be brittle when tested in realistic conditions. In this paper we approach the problem by training a neural MT system to learn how to use custom terminology when provided with the input. Comparative experiments show that our method is not only more effective than a state-of-the-art implementation of constrained decoding, but is also as fast as constraint-free decoding.