WTS 2024/Submissions/Ambiguous Pronoun Resolution on Wikipedia Texts

ID :
Author(s): Adith John Rajeev, Rahothvarman P, Kaveri Anuranjana	Username(s):	Type of submission: lightning talk
	Affiliation: IIIT Hyderabad	Theme(s): Technology
Abstract: Anaphora resolution is essential for NLU and machine translation but is challenging for low-resource languages like the Dravidian languages due to very little annotated data. To address this, we create mGAP from the English GAP dataset, which was sourced from wiki texts. This noisy but valuable dataset supports the development of multilingual models for anaphora resolution. We aim to leverage mGAP to train models that not only address anaphora resolution in Dravidian languages but also enhance cross-lingual transfer learning, thereby improving multilingual capabilities and effective anaphora resolution of the models across these languages.		Slides: Slides will be uploaded to the organizers
		Level of advancement: advanced
		Special requirements:

How will this session be beneficial for the communities ? The research addresses the challenge of improving language processing for diverse languages, including those with limited resources. These methods could enhance content accuracy and coherence across languages on Wikipedia, and could therefore improve the quality and accessibility of information on the platform for underrepresented langauges, which supports Wikipedia's mission to provide comprehensive knowledge globally.

(register below and ask your questions now to the session organiser)