E-book reading applications and devices include functionality for bookmarking a page within an e-book to mimic physical bookmarks used with printed books. However, current e-book bookmarking features require the user to provide explicit manual input to bookmark a particular page and to learn and remember to use the bookmarking feature. This disclosure describes techniques for automatically bookmarking pages of potential interest within an e-book. With appropriate permissions from the user, page locations for the automatic bookmarks can be identified based on user engagement with the book content determined by inferring the reader’s intent and attention when reading the e-book based on user-permitted interaction data. Parameters related to the user’s engagement with the e-book content over time are encoded as a feature vector and converted to an adaptive histogram of attention scores. Additionally, the e-book contents are parsed with a suitable language model to obtain the e-book structure. The attention scores and e-book structure are combined by applying heuristics to generate high-quality bookmarks.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Shin, D, "Automatically Generating E-book Bookmarks from User Engagement", Technical Disclosure Commons, (January 30, 2023)