Maunendra Sankar Desarkar

I am an Associate Professor at the Computer Science and Engineering Department at IIT Hyderabad, India. Currently, I am also serving as the Head of the Department of the AI Department at IIT Hyderabad. My main research areas are Natural Language Processing, Information Retrieval, and Machine Learning. I am a part of the Natural Language and Information Processing (NLIP) Research Group at the Department of CSE, IIT Hyderabad. Prior to Joining IIT Hyderabad, I have worked for Samsung Research India Bangalore and Sybase Inc. (a SAP Company) in the past.

Recent News

  • Two papers accepted at PAKDD 2024. One is on non-toxic autosuggest generation - A joint work with Manish Gipta (Microsoft). This is the first publication from Aishwarya's PhD work. The other paper is from Debolena's PhD work on Image captioning. Again, first publication. Congratulations, Aishwarya and Debolena.
  • We are working on improving zero-short MT for extremely low resource languages, using noise injection techniques. Two papers on this theme are acceptted, in EMNLP (Findings) and EACL. congratulations Kaushal. Also, congratulations to Maharaj on his first publication from PhD.
  • Suvodip's work on dialog system evaluation is nominated for best short paper award in SIGDIAL 2023. Although it became the runner-up, it is an amazing feat and recognition from the dialog system community.
  • Our MAPG (Microsoft Academic Partnership Grant) work on Autosuggest generation is accepted in ECML-PKDD journal track.
  • Suvodip's work on masking for explainable dialog response evaluation is accepted in SIGDIAL 2023.
  • Sharan's work on unsupervised style transfer using masked language models is accepted in INLG 2023
  • Arkadipta's work on VisioTextual Attention for Grounded Cross-Lingual NLI is accepted in Elsevier Natural Language journal.
  • "Time-to-Event Modeling with Hypernetwork based Hawkes Process" - work from Manisha's PhD work is accepted in ACM SIGKDD 2023.
  • We have a publication in ACL 2023, on the interesting task of Diverse Headline Generation. Congratultions to all the contributors, from different batches and programs.