CS 6804 
Natural Language Processing Techniques for Applications on Noisy Data 

Spring 2016 
Alla Rozovskaya

Today, a wealth of data is available at our fingertips that we would
like to make use of to solve societal challenges in various domains. In
the healthcare domain, for example, we have access to large data sets
of electronic health records. In the educational domain, we have large
repositories of data generated from educational software related to
people's learning activities in educational settings. In the social
media domain, there are a lot of things that can be studied. We can
make predictions about disasters and epidemics; we can do sentiment and
opinion mining. To make use of this unstructured data, however, we need
NLP technology. Unfortunately, traditional NLP technology breaks down
on this noisy data, because it is trained to deal with well-formed text.

In this graduate seminar, we will read and discuss papers that
are concerned with developing computational models to process noisy
non-standard data. The students will also work on research projects to
develop techniques and approaches in this area.