SAMARTH SINHA

samarth [dot] sinha [at] mail [dot] utoronto [dot] ca

I am currently an AI Resident at Meta AI Research in London working with David Novotny and Andrea Vedaldi on 3D reconstruction. 


Previously, I finished my undergrad in Computer Engineering at the University of Toronto.  I was a visiting student at MILA working closely with Hugo Larochelle and Yoshua Bengio; I was also a visitor at UC Berkeley, working with Trevor Darrell and Zeynep Akata.

​

UPDATE: I am actively looking for 1-2 undergrad students to work with each semester (ideally longer). If you are interested in 3D reconstruction and NeRFs, then please send an email! 

Twitter
Google Scholar
IMG_4728.JPG
 

RESEARCH INTERESTS

My current research interests deals with 3D vision, neural rendering and 3D reconstruction from videos. I am generally interested in how to train better computer vision models.

​​

I have had the privilege to work with some amazing collaborators, and I am always looking for students and collaborations. If you are interested in anything on the website, please get in touch!

Get in Touch
 

RECENT NEWS

  • [June 2022] Paper at L4DC 2022 was named a Best Paper Finalist! 🥳

  • [May 2022] One paper accepted at ICML 2022 and one paper accepted to UAI 2022 as posters! 🧠

  • [Mar 2022] One paper accepted at L4DC 2022 and one paper accepted at CVPR 2022 both as Oral!🤖 👀

  • [Oct 2021] We are organising a ICCV 2021 social for students looking to get involved in research! More info here!

  • [Sept 2021] Two papers accepted to NeurIPS 2021! 🧠

  • [Sept 2021] One paper accepted to CoRL 2021! 🤖

  • [Sept 2021] Started my AI Residency at Facebook London! 📝

  • [July 2021] One paper accepted to IROS 2021! 🤖

  • [June 2021] Finished undergrad at Toronto 🎓

  • [May 2021] Started internship at Amazon Research! 📝

  • [Dec 2020] One paper accepted to AAAI 2021! 🧠

  • [Sept 2020] Two papers accepted to NeurIPS 2020, including a Spotlight! 🧠

  • [July 2020] One paper accepted to ECCV 2020! 👀

  • [May 2020] Stared as a student researcher at the Vector Institute 📝

  • [April 2020] Two papers accepted to ICML 2020! 🧠

  • [June 2019] One paper accepted to ICCV 2019 as an Oral! 👀

  • [May 2019] Started as a visiting student at MILA 📝

  • [March 2019] One paper accepted to CVPR 2019! 👀

  • [May 2018] Started as a visiting student at UC Berkeley 

🧠= ML conference; 👀 = CV conference; 🤖 = Robotics conference; 

 

PUBLICATIONS

KEYTR: KEYPOINT TRANSPORTER FOR 3D RECONSTRUCTION
OF DEFORMABLE OBJECTS IN VIDEOS

David Novotny, Ignacio Rocco, Samarth Sinha, Alexandre Carlier, Gael Kerchenbaum, Roman Shapovalov, Nikita Smetanin, Natalia Neverova, Ben Graham, Andrea Vedaldi

CVPR 2022 (Oral)

KOOPMAN Q-LEARNING: OFFLINE REINFORCEMENT LEARNING VIA SYMMETRIES OF DYNAMICS

Matthias Weissenbacher, Samarth Sinha, Animesh Garg, Yoshinobu Kawahara

ICML 2022 (poster)

STACKMIX: A COMPLIMENTARY MIX ALGORITHM

John Chen, Samarth Sinha, Anastasios Kyrillidis 

UAI 2022 (Poster)

EXPERIENCE REPLAY WITH LIKELIHOOD-FREE IMPORTANCE WEIGHTS

Samarth Sinha*, Jiaming Song*, Animesh Garg, Stefano Ermon

L4DC 2022 (Oral)

CONSISTENCY REGULARIZATION FOR VARIATIONAL AUTOENCODERS

Samarth Sinha, Adji B. Dieng
NeurIPS 2021 (Poster)

CHARACTERIZING GENERALIZATION UNDER OUT-OF-DISTRIBUTION SHIFTS IN DEEP METRIC LEARNING

Karsten Roth*, Timo Milbich*, Samarth Sinha, Ludwig Schmidt, Marzyeh Ghassemi, Bjorn Ommer

NeurIPS 2021 (Poster)

S4RL: SURPRISINGLY SIMPLE SELF-SUPERVISION FOR REINFORCEMENT LEARNING

Samarth Sinha, Ajay Mandlekar, Animesh Garg

CoRL 2021 (Poster)

LEARNING BY WATCHING: PHYSICAL IMITATION OF MANIPULATION SKILLS FROM HUMAN VIDEOS

Haoyu Xiong, Quanzhou Li, Yun-Chun Chen, Homanga Bharadhwaj, Samarth Sinha, Animesh Garg

IROS 2021 (Poster)

DIBS: DIVERSITY INDUCING INFORMATION BOTTLENECK IN MODEL ENSEMBLES

Samarth Sinha*, Homanga Bharadhwaj*, Anirudh Goyal, Hugo Larochelle, Animesh Garg, Florian Shkurti

AAAI 2021 (Poster)

CURRICULUM BY SMOOTHING

Samarth Sinha, Animesh Garg, Hugo Larochelle 

NeurIPS 2020 (Spotlight)

TOP-K TRAINING OF GANS: IMPROVING GAN PERFORMANCE BY THROWING AWAY BAD SAMPLES

Samarth Sinha*, Zhengli Zhao*, Anirudh Goyal, Colin Raffel, Augustus Odena 

NeurIPS 2020 (Poster)

D2RL: DEEP DENSE ARCHITECTURE FOR REINFORCEMENT LEARNING

Samarth Sinha*, Homanga Bharadhwaj*, Aravind Srinivas, Animesh Garg

NeurIPS Deep RL Workshop 2020

DIVA: DIVERSE VISUAL FEATURE AGGREGATION FOR DEEP METRIC LEARNING

Karsten Roth*, Timo Milbich*, Homanga Bharadhwaj, Samarth Sinha, Joseph Paul Cohen, Bjorn Ommer, Yoshua Bengio

ECCV 2020 (Poster)

REVISITING TRAINING STRATEGIES AND GENERALIZATION PERFORMANCE IN DEEP METRIC LEARNING

Karsten Roth*, Timo Milbich*, Samarth Sinha, Prateek Gupta, Bjoern Ommer, Joseph Paul Cohen

ICML 2020 (Poster)

SMALL-GAN: SPEEDING UP GAN TRAINING USING CORE-SETS

Samarth Sinha, Han Zhang, Anirudh Goyal, Yoshua Bengio, Hugo Larochelle, Augustus Odena 
ICML 2020 (Poster)

VARIATIONAL ADVERSARIAL ACTIVE LEARNING

Samarth Sinha*, Sayna Ebrahimi*, Trevor Darrell

ICCV 2019 (Oral)

GENERALIZED ZERO- AND FEW-SHOT LEARNING VIA ALIGNED VARIATIONAL AUTOENCODERS

Edgar Schönfeld, Sayna Ebrahimi, Samarth Sinha, Trevor Darrell, Zeynep Akata 

CVPR 2019 (Poster)