[Ltg] LTG Seminar [CLT Honours Students, June 7]

[Stephen.Wan at csiro.au]

----
LTG Seminar
 - see: http://www.clt.mq.edu.au/Events/Seminars.html

Monday, June 7, 2004 at 11am
Macquarie Uni, E6A 357
----

Speakers: CLT Honours Students: Hogan Ho, Jason Barles, James Ballantine

---
Speaker: Hogan Ho

Title: Error Correction of OCR-generated Texts Using Corpus-Specific
Information

Abstract:
Spelling error correction and the improvement of optical character
recognition (OCR) are independent areas that have been long researched.
The combination of these, namely post-OCR spelling error correction, has
received relatively less attention. The work carried out in my project
incorporates the use of corpus-specific information to help aid the
process of post-OCR spelling error correction.
---
Speaker: Jason Barles

Title: An Answer Comparison Engine for an Open-domain Intelligent
Tutoring System.

Abstract:
In a classroom environment students learn as a group and do not
necessarily receive the attention they need. An Intelligent Tutoring
System (ITS) is an educational computer program that provides
individualised tutoring for a student by following the methods of
one-on-one human tutors.

In my talk I will talk about an ITS called Tutor2U which has a
conversational interface that handles short answer questions in all
kinds of domains, while existing ITSs are created for specific domains.
The aim of the project is to create an answer comparison engine that is
capable of matching a student's answer, which can have syntactic and
semantic variation, to the gold standard expected answer. This answer
comparison engine obtains enough information from the student's answer
to facilitate a tailored response to the student. Techniques have been
borrowed from other ITSs as well as Question-Answering Systems to create
the answer comparison engine.
---
Speaker: James Ballantine

Title: Topic Segmentation in Spoken Dialogue

Abstract:
Topic segmentation is the division of language data into semantically 
coherent blocks, based on the topics they cover. Traditional topic 
segmentation techniques focus on written text; spoken dialogue presents 
different challenges such as lack of paragraph markings, different 
lexical topic-change cues, and lack of organised structure.

I will first describe the implementation and evaluation of existing 
techniques for the special case of transcripts of spoken dialogue. 
Secondly, I will discuss experiments in improving topic segmentation 
results for the domain of spoken dialogue.
---