Computer Science Seminar
Date: Monday, 19 February 2024
Time: 2:00 to 3:00 PM
Venue: Seminar Hall
Sanskrit Computational Linguistics: Tools, Annotation and Knowledge Graphs
Hrishikesh Terdalkar
IIT Kanpur.
19-02-24

---

Abstract

This presentation explores the computational facets of Sanskrit, focusing on intriguing applications within its literature, question answering, and manual annotation processes. Despite the richness of its classical literature, Sanskrit remains a computationally low resourced language. We discuss three strategies employed to target this problem: (1) creating engaging computational interfaces, (2) developing tools to assist researchers, and (3) establishing a foundational manual annotation process for dataset generation.

The exploration begins with an examination of various computational aspects related to the language and literature, including ancient numeral systems, techniques for maintaining textual correctness in verbal knowledge transmission, and a system of Sanskrit prosody which may be used for automatic text correction.

With the overarching goal of natural language question-answering, we delve into the construction of knowledge graphs using rule-based approaches and manual annotation. We explore the structure, features and applications of annotation tools. We conclude with a discussion on the future directions for Sanskrit Computational Linguistics.