PhD Dissertation Proposal by NGUYEN Thanh Son | Entity Summarization of Reviews and Micro-Reviews

About The Talk

Along with the regular review content, there is a new type of user-generated content arising from the prevalence of mobile devices and social media, that is Micro-reviews. Micro-reviews are bite-size reviews (usually under 200 characters), commonly posted on social media or check-in services, using a mobile device. They capture the immediate reaction of users, and they are rich in information, concise, and to the point. Both reviews and micro-reviews are useful for users to get to know the entity of interest, thus facilitating users in making their decision of purchasing or dining. However, the abundant number of both reviews and micro-reviews makes it increasingly difficult to go through them and extract the useful information. In this dissertation proposal, we propose to summarize reviews and micro-reviews to ease users in understanding entity (or a set of entities). We focus on two different scenarios when generating a summary: summarization for a single entity and summarization for a set of entities. For single-entity summarization, we propose a novel mining problem, which brings together the two disparate sources of review content for summarizing micro-reviews of an entity. The summaries are generated based on reviews' content as reviews are often coherent, well-written pieces of text, produced by an author who seeks to comprehensively describe her experience with the entity. For generating the summaries, we choose to either select a small number of reviews or synthesize a review. We perform thorough evaluations of our methodologies on real-life data collected from Foursquare and Yelp. For multi-entity summarization, we focus on summarizing for a set of entities. We address the problem of summarizing the micro-reviews of multiple entities in a collection by synthesizing new micro-reviews that pertain to the collection, rather than to the individual entities per se. We formulate this problem in terms of first finding a representation of the collection, by identifying a number of "aspects" that link common threads across two or more entities within the collection. We then synthesize a summary micro-review for each aspect. Our approach performs well on collections of Foursquare entities based on localities and categories, producing more representative and diverse summaries than the baselines.

Speaker Biography

NGUYEN Thanh Son is a fifth-year PhD candidate in Information Systems. He started his PhD programme in August 2013 under the supervision of Assistant Professor Hady W. Lauw. His research interests involve mining entity representation based on text data, algorithmic data mining, mining opinionated text, and entity ranking. He was awarded the SMU Presidential Doctoral Fellowship in 2015. From August 2016 to May 2017, he participated in a 10-month training residency at Carnegie Mellon University. After that, he went to IBM Research Lab in Dublin, Ireland for a 4-month internship. He obtained his Bachelor Degree from University of Engineering and Technology (UET), Vietnam National University, Hanoi (VNU) in 2010.