Bilingual Learning of Multi-sense Embeddings with Discrete Autoencoders

摘要: We present an approach to learning multi-sense word embeddings relying both on monolingual and bilingual information. Our model consists of encoder, which uses context (i.e. a parallel sentence) choose sense for given word, decoder predicts words based the chosen sense. The two components are estimated jointly. observe that representations induced from data outperform counterparts across range evaluation tasks, even though crosslingual information is not available at test time.