ANKUR SIKARWAR
I am a master’s student in computer science at Mila and Université de Montréal, supervised by Prof. Aishwarya Agrawal and planning to advance to a PhD. Previously, I was a Research Engineer at Agency for Science, Technology and Research (A*STAR) Singapore, where I was supervised by Prof. Mengmi Zhang. I also spent a year at Microsoft Research India working with Dr. Navin Goyal. Prior to that, I was a Research Assistant at Kreiman Lab, Harvard University, where I was advised by Prof. Gabriel Kreiman.
6666 Saint-Urbain Street
Montreal, Quebec
H2S 3H1
news
Aug 2024
Dec 2023
Aug 2023
Joined Mila - Quebec AI Institute as a Graduate Research Assistant
Presented our paper on working memory in AI and Humans at NeurIPS 2023 in New Orleans
Awarded Fujitsu Laboratories Fellowship for MIT-Harvard Centre for Brains, Minds and Machines 2023 Summer Program
Aug 2023
July 2023
Dec 2022
Oct 2022
Attended MIT-Harvard Center for Brains, Minds and Machines 2023 summer program
Our paper on evaluating curricula in multi-task learning was accepted at ICCV 2023
Presented our paper on Grounded Compositional Generalization in Transformers at EMNLP 2022 in Abu Dhabi
Joined Deep NeuroCognition Lab, A*STAR Singapore as a Research Engineer
experience
Mila - Quebec AI Institute
Aug 2024 - Present
Graduate Research Student | Advisor : Prof. Aishwarya Agrawal
Working on text-to-image generation using Chain-of-Thought reasoning to improve the model’s ability to follow complex, multi-step instructions. Also, working on understanding training dynamics that lead to better compositional generalization.
Oct 2022 - Jun 2024
Deep NeuroCognition Lab | Agency for Science, Technology and Research
Research Engineer | Advisor : Prof. Mengmi Zhang
Developing robust memory-augmented networks for out-of-distribution generalization. Also, working on self-supervised methods for contextual reasoning.
MIT Center for Brains, Minds and Machines
Summer Course Fellow
Summer of 2023
Awarded Fujitsu Laboratories Fellowship for MIT Brains, Minds, and Machines Summer Program.
July 2021 - Aug 2022
Microsoft Research
Research Intern | Advisor : Dr. Navin Goyal
Developed models capable of generalizing compositionally in grounded language understanding tasks. Also, worked on the mechanistic interpretability of grounding and composition in multimodal transformers.
Jan 2021 - July 2021
Kreiman Lab | Harvard University
Research Assistant | Advisor : Prof. Gabriel Kreiman
Investigated the efficacy of cross-modal attention in vision-language tasks. Demonstrated that attention in transformers correlates more with human attention when compared with CNN/LSTM networks.
May 2019 - July 2019
Center for Visual Information Technology | IIIT Hyderabad
Research Intern | Advisor : Prof. Avinash Sharma
Worked on an end-to-end network for reconstructing 3D models of humans from monocular video. Developed tools for pre-processing & generating 3D mesh data of humans from a vertex-based template model.
July 2017 - June 2021
Birla Institute of Technology, Mesra
Undergraduate Researcher | Advisor : Prof. Swati Prasad
Developed interpretable and modular models for visual question answering via disentangled representation of visual scenes. Explored beam search algorithm for generating structured programs from natural language questions.
academic service
Reviewer
Organizer
NeurIPS’23, EMNLP’23, ACL’23, EMNLP’22
Reading Group, Deep NeuroCognition Lab, A*STAR Singapore