Improving Multilingual Sentence Embedding using Bi-directional Dual Encoder with Additive Margin Softmax

Abstract

In this paper, we present an approach to learn multilingual sentence embeddings using a bi-directional dual-encoder with additive margin softmax. The embeddings are able to achieve state-of-the-art results on the United Nations (UN) parallel corpus retrieval task. In all the languages tested, the system achieves P@1 of 86% or higher. We use pairs retrieved by our approach to train NMT models that achieve similar performance to models trained on gold pairs. We explore simple document-level embeddings constructed by averaging our sentence embeddings. On the UN document-level retrieval task, document embeddings achieve around 97% on P@1 for all experimented language pairs. Lastly, we evaluate the proposed model on the BUCC mining task. The learned embeddings with raw cosine similarity scores achieve competitive results compared to current state-of-the-art models, and with a second-stage scorer we achieve a new state-of-the-art level on this task.

Accepted by IJCAI'19(International Joint Conference on Artificial Intelligence)

References in corpus (6)

• Achieving Human Parity on Automatic Chinese to English News Translation
• Attentive Pooling Networks
• ABCNN: Attention-Based Convolutional Neural Network for Modeling Sentence Pairs
• A Deep Neural Network Approach To Parallel Sentence Extraction
• Margin-based Parallel Corpus Mining with Multilingual Sentence Embeddings
• Filtering and Mining Parallel Data in a Joint Multilingual Space

Cited by in corpus (3)

• ComStreamClust: a communicative multi-agent approach to text clustering in streaming data
• ABSent: Cross-Lingual Sentence Representation Mapping with Bidirectional GANs
• TransAug: Translate as Augmentation for Sentence Embeddings