Aniket Rege

Hello There!

prof_pic.jpg

PhD Student, University

of Wisconsin-Madison

I’m a Ph.D student in Computer Sciences at the University of Wisconsin-Madison, advised by Prof. Ramya Vinayak working on human preference learning and generative models.

Previously, I was an masters student at the University of Washington, Seattle where I worked with the RAIVN Lab advised by Prof. Ali Farhadi and mentored by Aditya Kusupati. My MS research focused on large-scale efficient and deployable machine learning, visual representation learning, and web-scale search.

In Summer ‘22 I interned in the deep learning for autonomous vehicles team at NVIDIA Redmond on Multiview LidarNet, supervised by Dr. Nikolai Smolyanskiy and Dr. Stan Birchfield. Previously, I worked for several years as a research engineer at Samsung Research India, working on perception problems for the AI Camera, including as a founding member of Single Take Camera, the USP of the Galaxy S20 flagship smartphone.

I am also generally interested in contrastive learning, learning from minimal supervision, and model robustness to real-world distribution shifts.

Reach me via aniketr[at]cs[dot]wisc[dot]edu

news

Feb 7, 2024 I published a beginner-friendly blog about MRL following OpenAI’s training their new Matryoshka embedding models!
Aug 21, 2023 AdANNS is accepted at NeurIPS 2023! A previous version was also accepted at ICLR ‘23 PML4DC Workshop.
May 31, 2023 I successfully defended my masters thesis!
Feb 7, 2023 Our contribution to Google’s large-scale efficient deep learning published as a Google Blog.
Sep 14, 2022 Matryoshka Representation Learning has been accepted for publication.
  1. NeurIPS 2022 (main conference and workshops: SSL and VTTA)
  2. ECCV 2022 CVinW workshop

selected publications

  1. search.gif
    AdANNS: A Framework for Adaptive Semantic Search
    In Advances in Neural Information Processing Systems (NeurIPS), 2023
  2. matryoshka.gif
    Matryoshka Representation Learning
    In Advances in Neural Information Processing Systems (NeurIPS), 2022