Subsentence Extraction from Text Using Coverage-Based Deep Learning Language Models

Sentiment prediction remains a challenging and unresolved task in various research fields, including psychology, neuroscience, computer science. This stems from its high degree of subjectivity limited input sources that can effectively capture the actual sentiment. be even more with only text-based input. Meanwhile, rise deep learning an unprecedented large volume data have paved way for artificial intelligence to perform impressively accurate predictions or human-level reasoning. Drawing inspiration this, we propose coverage-based sentiment subsentence extraction system estimates span text recursively feeds this information back networks. The predicted consists auxiliary expressing is important building block enabling vivid epic delivery (within scope paper) other natural language processing tasks such as summarisation Q&A. Our approach outperforms state-of-the-art approaches by margin (i.e., Average Jaccard scores 0.72 0.89). For evaluation, designed rigorous experiments consisting 24 ablation studies. Finally, our learned lessons are returned community sharing software packages public dataset reproduce results presented paper.