We propose a generative model for text generation, which exhibits disentangled latent representations of syntax and semantics. Contrary to previous work, this model does not need syntactic information such as constituency parses, or semantic information such as paraphrase pairs. Our model relies solely on the inductive bias found in attention-based architectures such as Transformers. In the attention of Transformers, keys handle information selection while values specify what information is conveyed. Our model, dubbed QKVAE, uses Attention in its decoder to read latent variables where one latent variable infers keys while another infers values. We run experiments on latent representations and experiments on syntax/semantics transfer which show that QKVAE displays clear signs of disentangled syntax and semantics. We also show that our model displays competitive syntax transfer capabilities when compared to supervised models and that comparable supervised models need a fairly large amount of data (more than 50K samples) to outperform it on both syntactic and semantic transfer. The code for our experiments is publicly available.
Semi-Supervised Variational Autoencoders (SSVAEs) are widely used models for data efficient learning. In this paper, we question the adequacy of the standard design of sequence SSVAEs for the task of text classification as we exhibit two sources of overcomplexity for which we provide simplifications. These simplifications to SSVAEs preserve their theoretical soundness while providing a number of practical advantages in the semi-supervised setup where the result of training is a text classifier. These simplifications are the removal of (i) the Kullback-Liebler divergence from its objective and (ii) the fully unobserved latent variable from its probabilistic model. These changes relieve users from choosing a prior for their latent variables, make the model smaller and faster, and allow for a better flow of information into the latent variables. We compare the simplified versions to standard SSVAEs on 4 text classification tasks. On top of the above-mentioned simplification, experiments show a speed-up of 26%, while keeping equivalent classification scores. The code to reproduce our experiments is public.
Cet article présente notre participation à l’édition 2020 du Défi Fouille de Textes DEFT 2020 et plus précisément aux deux tâches ayant trait à la similarité entre phrases. Dans notre travail nous nous sommes intéressé à deux questions : celle du choix de la mesure du similarité d’une part et celle du choix des opérandes sur lesquelles se porte la mesure de similarité. Nous avons notamment étudié la question de savoir s’il fallait utiliser des mots ou des chaînes de caractères (mots ou non-mots). Nous montrons d’une part que la similarité de Bray-Curtis peut être plus efficace et surtout plus stable que la similarité cosinus et d’autre part que le calcul de similarité sur des chaînes de caractères est plus efficace que le même calcul sur des mots.