Latent Variable Model for Multi-modal Translation

Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 6392–6405 Florence, Italy, July 28 - August 2, 2019. c⃝2019 Association for Computational Linguistics 6392 Latent Variable Model for Multi-modal Translation Iacer Calixto Miguel Rios ILLC The University of Amsterdam {iacer.calixto,m.riosgaona,w.aziz}@uva.nl Wilker Aziz Abstract In this work, we propose to model the in- teraction between visual and textual features for multi-modal neural machine translation (MMT) through a latent variable model. This latent variable can be seen as a multi-modal stochastic embedding of an image and its de- scription in a foreign language. It is used in a target-language decoder and also to pre- dict image features. Importantly, our model formulation utilises visual and textual inputs during training but does not require that im- ages be available at test time. We show that our latent variable MMT formulation im- proves considerably over strong baselines, in- cluding a multi-task learning approach (Elliott and K´ad´ar, 2017) and a conditional variational auto-encoder approach (Toyama et al., 2016). Finally, we show improvements due to (i) pre- dicting image features in addition to only con- ditioning on them, (ii) imposing a constraint on the KL term to promote models with non- negligible mutual information between inputs and latent variable, and (iii) by training on additional target-language image descriptions (i.e. synthetic data). 1 Introduction Multi-modal machine translation (MMT) is an exciting novel take on machine translation (MT) where we are interested in learning to translate sentences in the presence of visual input (mostly images). In the last three years there have been shared tasks (Specia et al., 2016; Elliott et al., 2017; Barrault et al., 2018) where many research groups proposed different techniques to integrate images into MT, e.g. Caglayan et al. (2017); Libovick´y and Helcl (2017). Most MMT models expand neural machine trans- lation (NMT) architectures (Sutskever et al., 2014; Bahdanau et al., 2015) to additionally condition on an image in order to compute the likelihood of a translation in context. This gives the model a chance to exploit correlations in visual and lan- guage data, but also means that images must be available at test time. An exception to this rule is the work of Toyama et al. (2016) who exploit the framework of conditional variational auto-encoders (CVAEs) (Sohn et al., 2015) to decouple the en- coder used for posterior inference at training time from the encoder used for generation at test time. Rather than conditioning on image features, the model of Elliott and K´ad´ar (2017) learns to rank image features using language data in a multi-task learning (MTL) framework, therefore sharing pa- rameters between a translation (generative) and a sentence-image ranking model (discriminative). This similarly exploits correlations between the two modalities and has the advantage that images are also not necessary at test time. In this work, we also aim at translating with- out images at test time, yet learning a visually grounded translation model. To that end, we re- sort to probabilistic modelling instead of multi-task learning and estimate a …

Content truncated to 500 words