THUMT: An Open Source Toolkit for Neural Machine Translation

Abstract

This paper introduces THUMT, an open-source toolkit for neural machine translation (NMT) developed by the Natural Language Processing Group at Tsinghua University. THUMT implements the standard attention-based encoder-decoder framework on top of Theano and supports three training criteria: maximum likelihood estimation, minimum risk training, and semi-supervised training. It features a visualization tool for displaying the relevance between hidden states in neural networks and contextual words, which helps to analyze the internal workings of NMT. Experiments on Chinese-English datasets show that THUMT using minimum risk training significantly outperforms GroundHog, a state-of-the-art toolkit for NMT.

4 pages, 1 figure

References in corpus (4)

• Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling
• ADADELTA: An Adaptive Learning Rate Method
• Is Neural Machine Translation Ready for Deployment? A Case Study on 30 Translation Directions
• Agreement-based Learning of Parallel Lexicons and Phrases from Non-Parallel Corpora

Cited by in corpus (26)

• Exploiting Deep Representations for Neural Machine Translation
• Multi-Head Attention with Disagreement Regularization
• Neutron: An Implementation of the Transformer Translation Model and its Variants
• Convolutional Self-Attention Networks
• On the Weaknesses of Reinforcement Learning for Neural Machine Translation
• Improving Back-Translation with Uncertainty-based Confidence Estimation
• Neural Machine Translation: A Review of Methods, Resources, and Tools
• Neural Machine Translation with Adequacy-Oriented Learning
• Self-Attention with Structural Position Representations
• MultiMWE: Building a Multi-lingual Multi-Word Expression (MWE) Parallel Corpora
• Shared-Private Bilingual Word Embeddings for Neural Machine Translation
• YANMTT: Yet Another Neural Machine Translation Toolkit
• Fast and Simple Mixture of Softmaxes with BPE and Hybrid-LightRNN for Language Generation
• Document-level Neural Machine Translation with Document Embeddings
• Neural Machine Translation: A Review and Survey
• Bilingual Mutual Information Based Adaptive Training for Neural Machine Translation
• Learning Contextualized Sentence Representations for Document-Level Neural Machine Translation
• Neuron Interaction Based Representation Composition for Neural Machine Translation
• Dynamic Context Selection for Document-level Neural Machine Translation via Reinforcement Learning
• Alternated Training with Synthetic and Authentic Data for Neural Machine Translation
• Document-level Neural Machine Translation with Associated Memory Network
• Modeling Bilingual Conversational Characteristics for Neural Chat Translation
• Towards Making the Most of Dialogue Characteristics for Neural Chat Translation
• Learning to Detect Unacceptable Machine Translations for Downstream Tasks
• Learning to Copy for Automatic Post-Editing
• Regularized Context Gates on Transformer for Machine Translation