INVITED TALKS

1. Prof Aravind K Joshi, University of Pennsylvania, USA
   
   Title of the talk: THE BOUNDARY BETWEEN SENTENCE AND DISCOURSE: ANNOTATION OF DISCOURSE CONNECTIVES AND THEIR ARGUMENTS.
   
   Abstract

   We will discuss some issues concerning the transition from sentence to discourse, in particular, those aspects of discourse structure that can be modeled in almost the same way as the sentence structure. Discourse connectives and their argument structures suggest a way of integrating this level of discourse structure with sentence structure. We will examine the similarities and differences between predicate argument structure at the sentence level and the discourse connectives and their arguments at the discourse level. We will also describe a project for annotating a corpus with discourse connectives and their arguments.

2. Prof Gerard Huet, INRIA - Rocquencourt, France
   
   Title of the talk: Towards Computational Processing of Sanskrit
   
   Abstract

   The talk will present the current state of the art of computational linguistics in the Sanskrit language. It will demonstrate a Sanskrit tagger, and discuss its limitations and possible improvements. It will discuss some on-going efforts at gathering comprehensive Sanskrit lexicons and digital corpus, and preliminary attempts at creating a reference tree-bank of common syntactic constructions. It will conclude by discussing foreseeable progress towards a Perseus-like Sanskrit Digital Library.

3. Prof Pushpak Bhattacharyya, IIT Bombay, India
   
   Title of the talk: Interlingua Based Machine Translation- a perspective
   
   Abstract: We present here some important developments in in the field of machine translation that led to the interlingua based approach to MT. After touching upon the criteria for choosing and designing an interlingua we discuss the language divergence issues that pose problems for interlingua based MT. In this context we present our work on interlingua based MT involving English, Hindi and Marathi and discuss the importance of lexical resources in this matter.