Preksha Nema Home Page
PhD Scholar ( Roll No: CS15D201 )
(Converted from M.Tech to M.Tech+PhD Dual-Degree)
Email : CS15D201 [at] smail [dot] iitm [dot] ac [dot] in   
Advisor(s) : Mitesh Khapra


Publications : (Last Five, while at IITM)View All

  • The Heads Hypothesis: A Unifying Statistical Approach Towards Understanding Multi-Headed Attention in BERT. 
    Authors : Madhura Pande, Aakriti Budhraja, Preksha Nema, Pratyush Kumar, Mitesh Khapra
    Appeared in Thirty-Fifth AAAI Conference on Artificial Intelligence, AAAI 2021, Thirty-Third Conference on Innovative Applications of Artificial Intelligence, IAAI 2021, The Eleventh Symposium on Educational Advances in Artificial Intelligence, EAAI 2021, Virtual Event, February 2-9, 2021, pp.13613-13621, Feb 2021
  • On the weak link between importance and prunability of attention heads. 
    Authors : Aakriti Budhraja, Madhura Pande, Preksha Nema, Pratyush Kumar, Mitesh Khapra
    Appeared in Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing, EMNLP 2020, Online, November 16-20, 2020 (EMNLP 2020) ,pp.3230-3235, Nov 2020
  • Towards Interpreting BERT for Reading Comprehension Based QA. 
    Authors : Sahana Ramnath, Preksha Nema, Deep Sahni, Mitesh Khapra
    Appeared in Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing, EMNLP 2020, Online, November 16-20, 2020 (EMNLP 2020) ,pp.3236-3242, Nov 2020
  • Towards Transparent and Explainable Attention Models. 
    Authors : Akash Kumar Mohankumar, Preksha Nema, Sharan Narasimhan, Mitesh Khapra, Balaji Vasan Srinivasan, Balaraman Ravindran
    Appeared in Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, ACL 2020, Online, July 5-10, 2020 (ACL 2020) ,pp.4206-4216, Jul 2020
  • Let's Ask Again: Refine Network for Automatic Question Generation. 
    Authors : Preksha Nema, Akash Kumar Mohankumar, Mitesh Khapra, Balaji Vasan Srinivasan, Balaraman Ravindran
    Appeared in Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing, EMNLP-IJCNLP 2019, Hong Kong, China, November 3-7, 2019, pp.3312-3321, Nov 2019

(Recent) Teaching Assistantship : View All  |  Back to top

Jan 2019 - May 2019 : - Deep Learning (CS7015)
Jul 2018 - Nov 2018 : - Introduction to Programming (CS1100)
Jan 2018 - May 2018 : - Deep Learning (CS7015)
Jul 2017 - Nov 2017 : - Topics in Deep Learning (CS7016)
Jul 2016 - Nov 2016 : - Machine Learning (CS5011)

Awards and Recognitions :View All  |  Back to top