Deep Chakraborty, Ph.D.

About Me

👋 Hi, I am Deep, a computer scientist and engineer. Currently, I work at Waymo.

I did my Ph.D. from the beautiful University of Massachusetts Amherst where I investigated fundamental problems in computer vision through the lens of self-supervised learning (SSL) and information theory, under the supervision of Erik Learned-Miller at the Vision lab. I also worked on applying SSL to remote sensing and thermal images for object discovery and detection.

My industry experience includes two research internships at Apple and one at Philips Lighting Research, where I worked on a variety of topics from scene understanding to audio processing.

In my free time, I love to experiment with coffee brewing techniques (inspired by James Hoffman), go on long rides on my road bike, and listen to classic rock music (I’m a big fan of Queen).

Research Interests

• Machine Learning: self-supervised learning, unsupervised learning, information theory
• Computer Vision: scene understanding, object detection, tracking

News

[July 2026] I joined Waymo as a Machine Learning Engineer in the Degraded Road Surfaces and Weather (DRAW) team.
[May 2026] I successfully defended my Ph.D. thesis titled “Understanding, Improving, and Applying Self-Supervised Neural Representations”.

Publications

2025 | 2024 | 2022 | 2019 | 2018 | 2016

^{[* = Authors Contributed Equally]}

2025

• Radial-VCReg: More Informative Representation Learning through Radial Gaussianization
  Yilun Kuang, Yash Dagade, Deep Chakraborty, Erik Learned-Miller, Randall Balestriero, Tim G. J. Rudner, Yann LeCun
  NeurIPS 2025 Workshop on Unifying Representations in Neural Models
  NeurIPS 2025 Workshop on Symmetry and Geometry in Neural Representations
  [PDF] [Code]

• A Survey on Data Curation for Visual Contrastive Learning: Why Crafting Effective Positive and Negative Pairs Matters
  Shasvat Desai, Debasmita Ghose, Deep Chakraborty
  In Submission
  [arXiv]

• Improving Pre-trained Self-Supervised Embeddings Through Effective Entropy Maximization
  Deep Chakraborty, Yann LeCun, Tim G. J. Rudner, Erik Learned-Miller
  International Conference on Artificial Intelligence and Statistics (AISTATS)
  [PDF] [arXiv] [Code]

2024

• Squeezing Water from a Stone: Improving Pre-trained Self-Supervised Embeddings Through Effective Entropy Maximization
  Deep Chakraborty, Tim G. J. Rudner, Erik Learned-Miller
  NeurIPS 2024 Workshop on Self-Supervised Learning - Theory and Practice
  [PDF]

2022

• Self-Supervised Learning to Guide Scientifically Relevant Terrain Categorization in Martian Terrain Images
  Tejas Panambur*, Deep Chakraborty*, Melissa Meyer, Ralph Milliken, Erik Learned-Miller, Mario Parente
  IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)
  [PDF] [arXiv] [Code]

2019

• Nonparallel emotional speech conversion
  Jian Gao, Deep Chakraborty, Hamidou Tembine, Olaitan Olaleye
  Annual Conference of the International Speech Communication Association (INTERSPEECH)
  [PDF] [arXiv] [Project Page]

• Pedestrian Detection in Thermal Images using Saliency Maps
  Debasmita Ghose*, Shasvat Desai*, Sneha Bhattacharya*, Deep Chakraborty*, Madalina Fiterau, Tauhidur Rahman
  IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)
  [PDF] [arXiv] [Project Page] [Spotlight Video]

2018

• Unsupervised Hard Example Mining from Videos for Object Detection
  SouYoung Jin*, Aruni RoyChowdhury*, Huaizu Jiang, Ashish Singh, Aditya Prasad, Deep Chakraborty, Erik Learned-Miller
  European Conference on Computer Vision (ECCV)
  [PDF] [arXiv] [Project Page]

2016

• Bird Call Identification using Dynamic Kernel based Support Vector Machines and Deep Neural Networks
  Deep Chakraborty, Paawan Mukker, Padmanabhan Rajan, Dileep A.D.
  IEEE International Conference on Machine Learning and Applications (ICMLA)
  [PDF]

Activities

• Reviewer for CVPR, ICCV, ECCV, BMVC, NeurIPS
• Speaker at EMBER
• Mentor at Eureka!

Powered by Jekyll and Minimal Light theme.