BERT4SO: Neural Sentence Ordering by Fine-tuning BERT
This work addresses sentence ordering for natural language processing applications, representing an incremental improvement over existing methods.
The paper tackled the sentence ordering problem by fine-tuning BERT with a novel method called BERT4SO, achieving state-of-the-art results on five benchmark datasets.
Sentence ordering aims to arrange the sentences of a given text in the correct order. Recent work frames it as a ranking problem and applies deep neural networks to it. In this work, we propose a new method, named BERT4SO, by fine-tuning BERT for sentence ordering. We concatenate all sentences and compute their representations by using multiple special tokens and carefully designed segment (interval) embeddings. The tokens across multiple sentences can attend to each other which greatly enhances their interactions. We also propose a margin-based listwise ranking loss based on ListMLE to facilitate the optimization process. Experimental results on five benchmark datasets demonstrate the effectiveness of our proposed method.