Targeted Sentiment Analysis (TSA) is a contemporary AI method for generating valuable insights from user reviews. Such insights may aid consumers in their decision-making, or help companies when they strive to understand customer satisfaction and guide marketing campaigns.
In this session, we will dive into recent advancements in TSA, including the first open-domain TSA benchmark, and a multi-domain TSA system that can process user reviews from diverse product and service domains. We will also cover a demo of a real-world TSA system, developed within IBM, which allows any user to use TSA through IBM's Watson Studio.
The demo will show how hundreds of reviews can be quickly analyzed with TSA, and how the pros and cons of a particular business may be easily extracted.
Presenter: Matan Orbach
Matan Orbach is a Research Staff Member at IBM Research AI. Since joining IBM in 2014, he has worked on a diverse set of NLP tasks, including, among others, multilingual stance detection and targeted sentiment analysis. Before that, Matan led a team within Project Debater, an IBM Grand Challenge, which focused on rebuttal generation through the use of principled arguments. Prior to joining IBM, Matan received his M.Sc. from the faculty of Electrical Engineering at the Technion, where his research focused on graph-based semi-supervised learning.
You might also want to check out the other events hosted through this Meetup Group at: https://www.meetup.com/technical-group-hosted-by-ibm/