Adaptive Summaries: A Personalized Concept-based Summarization Approach by Learning from Users’ Feedback

Big quantities of textual facts in our every day lives make computerized summarization a precious endeavor. Nevertheless, distinct end users could have distinct qualifications expertise and cognitive bias. Hence, it is impossible to deliver a summary that satisfies all end users.

A new research on arXiv.org proposes an interactive summarization system where by end users can select which information and facts they want to include things like.

Picture: Geralt by using pixabay.com, cost-free licence

Consumers select the duration of the summary and give feed-back in an iterative loop. They can pick or reject a thought, define the degree of importance, and give the self esteem degree. An integer linear optimization functionality maximizes user-centered content variety. Moreover, the suggested resource does not call for reference summaries for teaching. An empirical verification displays that working with users’ feed-back can help them to uncover the preferred information and facts.

Exploring the incredible volume

Read More