The accelerating decline in book reading among youth indicates an imminent issue with their mar- keting. If the current trend continues, literary read- ing will disappear in half a century. On the contrary, filmmaking has figured out various ways to grat- ify the audience instantly. We believe providing powerful quotations while previewing books can engage readers. Formally, given a book, we extract quotes from it that people could find interesting. We use Transformer models to identify potential quotes in a book, similar to reading comprehension and summarization tasks, and further prune the predictions with classification techniques. Initial results have corroborated that the language models can capture such intriguing lines in long-form documents
Rajasekhar Mekala rmekala@uci.edu
Agniraj Baikani abaikani@uci.edu
Sai Vineeth Doddala sdoddala@uci.edu