Semantically Driven Sentence Fusion: Modeling and Evaluation
This addresses a bottleneck in natural language generation for researchers and practitioners, though it is incremental as it builds on existing fusion methods.
The paper tackles the problem of limited robustness in sentence fusion models due to single-reference training and evaluation, and shows that expanding ground-truths into multiple references via connective phrase equivalence classes improves performance over state-of-the-art models.
Sentence fusion is the task of joining related sentences into coherent text. Current training and evaluation schemes for this task are based on single reference ground-truths and do not account for valid fusion variants. We show that this hinders models from robustly capturing the semantic relationship between input sentences. To alleviate this, we present an approach in which ground-truth solutions are automatically expanded into multiple references via curated equivalence classes of connective phrases. We apply this method to a large-scale dataset and use the augmented dataset for both model training and evaluation. To improve the learning of semantic representation using multiple references, we enrich the model with auxiliary discourse classification tasks under a multi-tasking framework. Our experiments highlight the improvements of our approach over state-of-the-art models.