Mei awarded NSF grant to refine language models

Assistant Professor Qiaozhu Mei has been awarded a two-year grant from the National Science Foundation to fund his project “Wordsmith in the Cloud: Refining Language Models Using Web-Scale Language Networks.”

The $214,985 award will fund the first attempt to refine language models with Web-scale language networks. The techniques will bridge the gap between cloud computing and text information management using language models.

Qiaozhu Mei

The developed techniques are also expected to solve challenging research problems, such as the construction of heterogeneous language networks from Web-scale corpora, the estimation of tie strengths in language networks, and the regularization of language models with multiple language networks.

Using the power of the Cloud, Qiaozhu and his team will be able to efficiently construct and manage the language networks from Web-scale corpora, steer the regularization framework and the refining processes with these Web-scale language networks, and apply the refined language models to real world text mining applications.

The project will produce large scale language networks from a variety of contexts, including general corpora like the Web, domain-specific corpora like scientific literature and healthcare, and community-specific corpora corresponding to the online social communities.

A Ph.D. student and two master’s students from UMSI will gain valuable interdisciplinary analytic training while working with Qiaozhu on this project.