3 May: 2 seminars on text and image -- Wednesday, May 8

Dublin Computational Linguistics Research Seminar: Index of May 2013 | Dublin Computational Linguistics Research Seminar - Index of year: 2013 | Full index


Natural Language Processing?

2.45-3.15pm Discussion

3.15-4.00 pm Khurshid Ahmad, Talking Pictures and Ontological Indices

4.00-4.30 Discussion and wrap-up


Talking pictures and seeing texts
A seminar on words and images
Place: Maxwell Lecture Theatre, Hamilton Building, Trinity College Dublin
Time: 2.00-4.30 pm, Wednesday 8th May, 2013.
Chair: Declan O’Sullivan, Associate Professor and Head, Intelligent Systems Discipline, TCD

This seminar is first in the series by the Intelligent Systems
Discipline, SCSS, Trinity College. The intention is to bring together
speakers in language, vision and graphics, and knowledge engineering
especially ontology, to address key questions in multi-modal
communications. This type of communication involves at least two
modes of human discourse –images, written word, speech, and other
symbol systems. Such an approach is critical for building intelligent
systems in the future. Information has traditionally been delivered
using image and texts, journal papers, especially in life sciences
comprise video sequences of animal/plant behaviour, and now newspapers
have video clips on their web sites. Written texts have collateral
images, audio streams have written transcripts, and streaming texts
are now part of visual images displayed on televisual broadcasts.

There is a premium on building lexica for natural language processing,
extracting ‘meaning’ from texts automatically, and for building lexica
for naming images automatically. Patrick Hanks is our first speaker
for the day. He is an eminent researcher, author of Lexical Analysis:
Norms and Exploitations (2013), and editor of two six-volume
collections in the Routledge Critical concepts series: Lexicology
(2009) and Metaphor and Figurative Language (2012) ; he has been the
editor-in-chief of some of the dictionaries published by the Oxford
University Press.

The second speaker is a local one, Khurshid Ahmad, Professor of
Computer Science at TCD. He will talk about images and collateral
texts. His talk will focus on the retrieval of images from image data
bases and image streams by using visual cues and with the help of
collateral texts. He has worked with forensic scientists, cell
biologists and museum curators, on automatically indexing images using
learning technologies like neural nets.


2.00pm Patrick Hanks: What should (and should not) be in a lexicon for Natural Language Processing?
2.45-3.15pm Discussion
3.15-4.00 pm Khurshid Ahmad, Talking Pictures and Ontological Indices
4.00-4.30 Discussion and wrap-up

Dublin Computational Linguistics Research Seminar - Index of May 2013 | Index of year: 2013 | Full index