Hyperbolic sentence representations for solving Textual Entailment
This work addresses textual entailment for natural language processing, but it is incremental as it applies hyperbolic spaces, a known approach for hierarchical data, to this task.
The paper tackled textual entailment by embedding sentences in hyperbolic spaces using the Poincare ball, and it consistently outperformed baselines on the SICK dataset while being second only to Order Embeddings on SNLI for binary classification.
Hyperbolic spaces have proven to be suitable for modeling data of hierarchical nature. As such we use the Poincare ball to embed sentences with the goal of proving how hyperbolic spaces can be used for solving Textual Entailment. To this end, apart from the standard datasets used for evaluating textual entailment, we developed two additional datasets. We evaluate against baselines of various backgrounds, including LSTMs, Order Embeddings and Euclidean Averaging, which comes as a natural counterpart to representing sentences into the Euclidean space. We consistently outperform the baselines on the SICK dataset and are second only to Order Embeddings on the SNLI dataset, for the binary classification version of the entailment task.