DATE:

Friday, March 15th, 2019

TIME:

3:30 PM

LOCATION:

GMCS-314

SPEAKER:

Yang Xu, Ph.D, Assistant Professor of Computer Science at San Diego State University.

ABSTRACT:

The exponential growth of data enables cognitive scientists and linguists to understand
language-based communication at a resolution not previously possible. Each little piece
of data in the large-scale corpora contains subtle information of human dialogue and
language, and provides a potential clue to understand the principles behind. How to put
these pieces together and achieve a holistic understanding of the multifaceted nature
of human dialogue and language under different scales is the center of my research.

In the first part of the talk, I will cover my work on using computational techniques,
such as topic segmentation algorithms and language models, to build a novel information-theoretic
model of human dialogue, which captures the convergence between interlocutors at higher-level
representations and potentially unifies existing theories of dialogue. In the second part,
I will talk about the application of this information-theoretic perspective of dialogue to
two real world scenarios: predicting the success of task-oriented collaborations; justifying
the effect of social power on language usage. In the third part, I will introduce my ongoing
work on studying long-term language change using vector-space representations of words.

HOST:

Tao Xie, Professor of Computer Science Department

DOWNLOAD:

PDF File