Upaya at ArabicNLU Shared-Task: Arabic Lexical Disambiguation using Large Language Models
Pawan Kumar Rajpoot, Ashvini Jindal, Ankur P. Parikh · 2024
Disambiguating a word's intended meaning (sense) in a given context is important in Natural Language Understanding (NLU).WSD aims to determine the correct sense of ambiguous words in context.At the same time, LMD (a WSD variation) focuses on disambiguating location mention.Both tasks are vital in Natural Language Processing (NLP) and information retrieval, as they help correctly interpret and extract information from text.Arabic version is further challenging because of its morphological richness, encompassing a complex interplay of roots, stems, and affixes.This paper describes our solutions to both tasks, employing Llama3 and Cohere-based models under Zero-Shot Learning and Re-Ranking, respectively.Both the shared tasks were part of the second Arabic Natural Language Processing Conference co-located with ACL 2024.Overall, we achieved 1st rank in the WSD task (accuracy 78%) and 2nd rank in the LMD task (MRR@1 0.59).