Research‎ > ‎

Natural Langauge Processing

After my PhD, I shifted from my focus on AI from symbolic AI to a very application-oriented area, Natural Language Processing (NLP). My work on NLP is about analyzing text using semantics and common-sense knowledge. I am particularly interested in review data, e.g., reviews for products on Amazon. Recently, I gained interests in word embedding for disambiguation.

Customer Review Analysis

The Internet empowers people to easily leave their opinions to many things, merchandise, services, restaurants, places, apps, etc. They grow exponentially. However, those reviews are too many to be utilized be peer customers or service providers. Therefore, AI kicks in to convert the massive amounts of text into information for human to directly use. For example, the program can pick a few representative reviews out of millions for humans to read. My interest is how to make use of the experience and knowledge expressed in such a gold mine.


Automated summarization

Human beings, by our nature, are impatient animals. We like elevator pitches. We want condensed, important, and interesting information. Automated summarization is for this purpose.