Skip to content

I’m an Applied Scientist at Amazon AI, Palo Alto. Previously, I was a PhD student at the ILCC, University of Edinburgh supervised by Prof. Mirella Lapata and Prof. Frank Keller.

My thesis work focuses on using visual information to disambiguate meaning and learn representations across languages. My general interests include: image/video description generation, video to text summarization and multilingual multimodal representation learning.

In 2018 I interned at Facebook AI Research with Marcus Rohrbach and Mike Lewis and worked on generating narrative video descriptions. In 2016 summer I interned at Mircosoft Research Redmond where I worked with Margaret Mitchell on generating image descriptions for blind and visually impaired users.

Before joining my PhD I obtained my masters from Dept. of Computing and Information systems, University of Melbourne as part of Erasmus Mundus Language and Communication Technologies Program and worked under the supervision of Prof. Tim Baldwin. I recieved my undergraduate from IIIT Hyderabad with honors in Data Mining affiliated with Data Science and Analytics Center.

Quick Links: CVGoogle Scholar ProfileTwitter


June 2019: Our paper Neural Word Decomposition Models for Abusive Language Detection is accepted at Abusive Language Workshop, ACL’19

May 2019: Our paper on Multimodal Abstractive Summarization is accepted at ACL’19

Feb 2019: I am co-organizing Representation Learning for NLP at ACL’19. Consider submitting your work (Deadline 26th April).

Jan 2019: I am co-organizing Shortcomings of Vision and Language Workshop at NAACL’19. Consider submitting your work (Dealine 27th February).

Dec 2018: Shruti presented our spotlight paper at ViGIL worksop.

Nov 2018: I have successfully defended my thesis.

Nov 2018: Our HowTo Videos to Text Summarization paper is accepted as spotlight paper at ViGIL workshop co-located with NIPS 2018

Nov 2018: Presented our paper Telling the Stories of Social Media Videos at EMNLP 2018

Oct 2018: VentureBeat article on our VideoStory Dataset