Unsupervised Scientific Abstract Segmentation with Normalized Mutual Information
This work addresses the challenge of segmenting scientific abstracts for researchers and automated systems, but it is incremental as it builds on existing segmentation methods with a novel optimization technique.
The paper tackled the problem of automatically segmenting scientific abstracts into premises and conclusions, particularly for non-structured abstracts where conclusion positions are uncertain, by using Normalized Mutual Information (NMI) in an unsupervised approach called GreedyCAS, which achieved the best performance across all evaluation metrics on non-structured abstracts and outperformed baselines on structured abstracts as measured by P_k.
The abstracts of scientific papers consist of premises and conclusions. Structured abstracts explicitly highlight the conclusion sentences, whereas non-structured abstracts may have conclusion sentences at uncertain positions. This implicit nature of conclusion positions makes the automatic segmentation of scientific abstracts into premises and conclusions a challenging task. In this work, we empirically explore using Normalized Mutual Information (NMI) for abstract segmentation. We consider each abstract as a recurrent cycle of sentences and place segmentation boundaries by greedily optimizing the NMI score between premises and conclusions. On non-structured abstracts, our proposed unsupervised approach GreedyCAS achieves the best performance across all evaluation metrics; on structured abstracts, GreedyCAS outperforms all baseline methods measured by $P_k$. The strong correlation of NMI to our evaluation metrics reveals the effectiveness of NMI for abstract segmentation.