Generating Contradictory, Neutral, and Entailing Sentences
This addresses the challenge of controlled sentence generation for NLP applications like textual entailment, though it appears incremental as it builds on existing latent space and adversarial methods.
The paper tackles the problem of generating sentences with specific logical relationships (contradictory, neutral, entailing) to an input sentence by modeling a conditional latent space and training with an adversarial objective. The results show improvements in quality and diversity of generated sentences, as measured by BLEU scores against actual sentences using state-of-the-art RTE models.
Learning distributed sentence representations remains an interesting problem in the field of Natural Language Processing (NLP). We want to learn a model that approximates the conditional latent space over the representations of a logical antecedent of the given statement. In our paper, we propose an approach to generating sentences, conditioned on an input sentence and a logical inference label. We do this by modeling the different possibilities for the output sentence as a distribution over the latent representation, which we train using an adversarial objective. We evaluate the model using two state-of-the-art models for the Recognizing Textual Entailment (RTE) task, and measure the BLEU scores against the actual sentences as a probe for the diversity of sentences produced by our model. The experiment results show that, given our framework, we have clear ways to improve the quality and diversity of generated sentences.