Seminars

View all Seminars  |  Download ICal for this event

Generalisable Language Learning for Human Information Needs

Series: Department Seminar

Speaker: Dr. Pradeep Dasigi, Allen Institute for AI, USA

Date/Time: Aug 22 08:00:00

Location: Online Seminar

Abstract:
Much progress has been made in the past decade towards building Natural Language Processing (NLP) systems that understand human languages. I will discuss in this talk how we can further improve these technologies to efficiently assist humans with retrieving, summarising, and drawing insights from the information they need. I will present my prior work on defining NLP tasks that reflect real information needs and argue for the necessity of application-oriented evaluation of language technologies. Since building specialised NLP systems can be expensive, much of my prior work was also focussed on generalisation of these systems across tasks and textual domains. I will briefly summarise that work and discuss some open research questions in understanding generalisation towards efficiently building usable information assistants.

This is an online talk. The Teams Meeting link for the talk is: Link

Speaker Bio:
Pradeep Dasigi is a Senior Research Scientist at the Allen Institute for AI, specialising in Natural Language Processing. His current research interests lie in robustness, generalisation, and efficient adaptation of language models. He has a PhD in Language Technologies from Carnegie Mellon University and a Masters degree in Computer Science from Columbia University. He frequently hosts episodes on the NLP Highlights podcast.

Host Faculty: R Govindarajan