Semantic Convolutional Neural Machine Translation Using AMR for English-Vietnamese

Semantic representation can help in enforcing meaning preservation and handling data sparsity of neural machine translation models. This paper presents an extension of the convolutional neural machine translation model to incorporate Abstract Meaning Representation as a kind of semantic representation to reduce language ambiguity or alleviate data sparseness problems. Evaluating on translating from English to Vietnamese with a low resource setting in the domain of TED talks, we obtain promising results in terms of both perplexity reductions and improved BLEU scores over the baseline method.