Trainable Greedy Decoding for Neural Machine Translation
This addresses a relatively neglected aspect of neural machine translation for researchers and practitioners, though it is incremental as it builds on existing decoding frameworks.
The paper tackles the problem of decoding in neural machine translation by proposing a trainable greedy decoder that can be optimized for arbitrary decoding objectives, showing it generates better translations with minimal computational overhead across four language pairs.
Recent research in neural machine translation has largely focused on two aspects; neural network architectures and end-to-end learning algorithms. The problem of decoding, however, has received relatively little attention from the research community. In this paper, we solely focus on the problem of decoding given a trained neural machine translation model. Instead of trying to build a new decoding algorithm for any specific decoding objective, we propose the idea of trainable decoding algorithm in which we train a decoding algorithm to find a translation that maximizes an arbitrary decoding objective. More specifically, we design an actor that observes and manipulates the hidden state of the neural machine translation decoder and propose to train it using a variant of deterministic policy gradient. We extensively evaluate the proposed algorithm using four language pairs and two decoding objectives and show that we can indeed train a trainable greedy decoder that generates a better translation (in terms of a target decoding objective) with minimal computational overhead.