Purva Tendulkar
purva at gatech dot edu

I am a second year computer science MS student at Georgia Tech, advised by Prof. Devi Parikh. I also collaborate with Ani Kembhavi at AI2. I earned my Bachelors in Computer Engineering in 2018 from College of Engineering Pune, India.

In the past couple of years, I have had the fortune to intern / conduct research at AiBee Inc., Palo Alto (Summer 2019) mentored by Chunhui Gu and Juan Carlos Niebles and; Cyber-Physical Systems Research Group, NTU Singapore (Summer 2017) mentored by Arvind Easwaran and Anupam Chattopadhyay, on a diverse set of topics - ranging from vision & language to cyber security.

In my free time, I play the Harmonium instrument. I am very passionate about music. I love to learn about different music styles and their histories. I am also a trained Bharatnatyam dancer.

Email  /  CV  /  Google Scholar  /  LinkedIn  /  Github  /  Twitter

profile photo

I am interested in the broad areas of Machine Learning, Computer Vision & Language and Reinforcement Learning. My primary research interest is in connecting humans and AI, i.e., developing personable AI that humans can like, trust, teach and learn from. My work so far has attempted to address tasks along these lines to make AI:

  • creative, so that they may assist humans and spur human creativity
  • trustworthy, so that predictions made by such systems can be explained, trusted and more reliably deployed
  • learn to learn like humans do, so that they can give insight into how humans learn and made more immersive

3DSP Feel The Music: Automatically Generating A Dance For An Input Song
Purva Tendulkar, Abhishek Das, Ani Kembhavi, Devi Parikh
Keywords: dance, creativity, music, human studies
ICCC, 2020 (Oral)
paper / talk / website / press coverage
3DSP SQuINTing at VQA Models: Interrogating VQA Models with Sub-Questions
Ramprasaath R. Selvaraju, Purva Tendulkar, Devi Parikh, Eric Horvitz, Marco Ribeiro, Besmira Nushi, Ece Kamar
Keywords: visual question answering, consistency, sub-questions, reasoning
CVPR, 2020 (Oral)
paper / talk / data
3DSP Trick or TReAT: Thematic Reinforcement for Artistic Typography
Purva Tendulkar, Kalpesh Krishna, Ramprasaath R. Selvaraju, Devi Parikh
Keywords: computer vision, unsupervised learning, creativity
ICCC, 2019 (Oral, Best Presentation Award)
paper / talk / slides pdf / google slides / demo / code
3DSP Trajectory-based event detection
Purva Tendulkar, Chunhui Gu, Juan Carlos Niebles, Sinisa Todorovic, Silvio Savarese
Keywords: deep recurrent neural networks, human trajectories, classification
3DSP Blind image dehazing using internal patch recurrence
Purva Tendulkar, Kalpesh Krishna, Yuval Bahat
Keywords: computer vision, dehazing
3DSP Safety and Security Analysis in Cyber-Physical Systems using Metro-II
Keywords: cyber security, stuxnet, metropolis
Purva Tendulkar, Anupam Chattopadhyay, Arvind Easwaran
3DSP EvalPro: A project of BodhiTree
Purva Tendulkar, Varsha Apte
Keywords: django, html, css, javascript

(Design and CSS courtesy: Jon Barron and Amlaan Bhoi)