ABSTRACT:

 Tables are used to store and retrieve structured information in the databases. They are used to communicate business statistics, scientific results, sports results, demographic data, etc. Although they can be accessed through standard APIs such as structured query language (SQL), most users are not familiar with it.

To close the human-table-human interface, there are proposing a suite of NLP models to access and retrieve data from databases using natural language. It includes text-to-SQL and table mining and summarization. The goal is to provide a reliable, domain-independent, and natural end to end interface for users to interact with structured tabular data.

BIO:

Caiming Xiong is the Managing Director of AI Research at Salesforce. He received his Ph.D. in the Department of Computer Science and Engineering, University at Buffalo, SUNY in 2014. During his career, he has published more than 100 papers in deep learning, natural language processing, computer vision, and theoretical machine learning. He has won an Outstanding Paper Award at ACL 2019. He served as an area chair at ACL, EMNLP, NLPCC, etc., and co-organized workshops at different AI conferences.

Hosted by Professor Quanquan Gu

Via Zoom Webinar