NLP Research Talk: Using Deep Learning for Urdu Language Processing

Mehreen Alam
Assistant Professor, School of Computing & Lead Scientist, UrduX Research Lab
National University of Computer and Emerging Sciences (NUCES), Islamabad, Pakistan
Location
Virtual
Organizer
Natural Language Processing M.S. program

Join us on Zoom: https://ucsc.zoom.us/j/94645898203?pwd=NTR4V0I4cllFTWs4ejZrRUxoZTBqdz09

Description: UCSC's Natural Language Processing Program invites you to join us for a research presentation by Mehreen Alam on September 23rd. Alam envisions a world where language is no more a barrier. Urdu remains technologically under-resourced despite being vast, morphologically rich, and used by over a 100 million people across the globe. In their presentation, Alam will cover various aspects of Urdu language: historical connections especially in the regional context, its comparison with two other regional languages Hindi and Persian, and the challenges faced in integrating Urdu in the technological fabric. They will also cover how far the technological landscape has advanced for Urdu, in particular the pre-trained models and transfer learning have contributed to its progress. Specifically, the problem areas Alam plans to discuss are research areas of neural machine translation, multi-lingual machine translation, machine transliteration, word embeddings, and conversational agents. The methods and techniques Alam has employed to address the aforementioned problems include lstms, seq2seq, transformers, bert, and m2m. Alam will discuss how their work has contributed to the body of knowledge and how these techniques have helped solve the aforementioned research areas. After the presentation, participants will have an opportunity to share questions with the presenter.