Unsupervised Summarization Techniques

Dr. Guy Feigenblat

Abstract

The abundance of unstructured information raises the need for automatic systems that can “condense” information from various documents into a shorter length, readable summary. Such summaries may further be required to cover a specific information need (e.g., summarizing web search results, question answering).
Various methods have been proposed for the query-focused summarization task. These methods can be categorized based on two main dimensions namely: extractive vs. abstractive and supervised vs. unsupervised.

In this talk we will focus on extractive, query focused, unsupervised mutli-document summarization techniques. We will survey some of the recent works and their applications in the industry in general and in IBM in particular. I will then describe a state of the art summarization approach that is being developed in our lab.
This talk is based on a work published in SIGIR 2017.

Speaker

Photo of Dr. Guy Feigenblat

Dr. Guy Feigenblat is a team leader at the Language and Retrieval group in IBM Research AI.
In his current position Guy mainly focuses on the development of automatic document summarization algorithms (query-based, extractive, abstractive) for various domains and use cases. Prior to that he was involved in developing cognitive bots that can express and predict human emotions. Guy was also leading a project that focused on indoor location analyses and how to predict customer’s behavior in virtual and physical venues.

Guy holds a Ph.D. in computer science from Bar-Ilan University, under the supervision of Prof. Ely Porat. In the academia his research focused on data structures, hash-functions, pattern matching and data mining.

Lecture languages

EnglishHebrew

Topics

AI / Automation

Duration options

1 hour

Travel/delivery options

In-countryOutside of country: Open for discussionRemote via video conference

Country

Israel

Lecture booking request

Thank you for your interest in hosting an IBM speaker. Please fill out the following form with as much detail as possible. An IBM representative will reach out to discuss your booking request. All guest lectures are subject to availability and agreements under this collaboration are not legally binding.