On the Linguistic Representational Power of Neural Machine Translation Models

Abstract

Despite the recent success of deep neural networks in various spheres of Artificial Intelligence, their interpretability remains a challenge. We analyze the representations learned by neural machine translation (NMT) models at various levels of granularity and evaluate their quality through relevant extrinsic properties. More precisely, we seek answers to the following questions: (i) How accurately is word-morphology captured within the learned representations? (ii) Do the representations capture long-range syntactic dependencies? and (iii) Do the representations capture lexical semantics? We conduct a thorough investigation along several parameters: (i) Which layers in the architecture capture each of these linguistic phenomena; (ii) How does the choice of translation unit (word, character, or sub-word unit) impact the linguistic properties captured by the underlying representations? (iii) Do the encoder and decoder learn differently and independently? iv) Do the representations learned from multilingual NMT models, capture same amount of linguistic information as their bilingual counterparts? Our data-driven, quantitative evaluation illuminates important aspects in NMT models and their ability to capture various linguistic phenomena. We show that deep NMT models, learn a non-trivial amount of linguistic information. Notable findings include the following observations: i) Word morphology is captured at the lower layers of the model; (ii) In contrast, lexical semantics or non-local syntactic and semantic dependencies are better represented at the higher layers of the model; (iii) Representations learned using characters are more informed about word morphology compared to those learned using sub-word units; and (iv) Representations learned from multilingual models are richer compared to the bilingual models.