Publications

2022

  1. Cold Diffusion: Inverting Arbitrary Image Transforms Without Noise
    Bansal, Arpit, Borgnia, Eitan, Chu, Hong-Min, Li, Jie S, Kazemi, Hamid, Huang, Furong, Goldblum, Micah, Geiping, Jonas, and Goldstein, Tom
    arXiv preprint arXiv:2208.09392 2022
  2. Certified Neural Network Watermarks with Randomized Smoothing
    Bansal, Arpit, Chiang, Ping-Yeh, Curry, Michael J, Jain, Rajiv, Wigington, Curtis, Manjunatha, Varun, Dickerson, John P, and Goldstein, Tom
    In International Conference on Machine Learning 2022
  3. Can Neural Nets Learn the Same Model Twice? Investigating Reproducibility and Double Descent from the Decision Boundary Perspective
    Somepalli, Gowthami, Fowl, Liam,  Bansal, Arpit, Yeh-Chiang, Ping, Dar, Yehuda, Baraniuk, Richard, Goldblum, Micah, and Goldstein, Tom
    In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition 2022
  4. End-to-end Algorithm Synthesis with Recurrent Networks: Logical Extrapolation Without Overthinking
    Bansal, Arpit, Schwarzschild, Avi, Borgnia, Eitan, Emam, Zeyad, Huang, Furong, Goldblum, Micah, and Goldstein, Tom
    arXiv preprint arXiv:2202.05826 2022
  5. Transfer Learning with Deep Tabular Models
    Levin, Roman, Cherepanova, Valeriia, Schwarzschild, Avi,  Bansal, Arpit, Bruss, C Bayan, Goldstein, Tom, Wilson, Andrew Gordon, and Goldblum, Micah
    arXiv preprint arXiv:2206.15306 2022

2021

  1. Preventing unauthorized use of proprietary data: Poisoning for secure dataset release
    Fowl, Liam, Chiang, Ping-yeh, Goldblum, Micah, Geiping, Jonas,  Bansal, Arpit, Czaja, Wojtek, and Goldstein, Tom
    arXiv preprint arXiv:2103.02683 2021
  2. MetaBalance: High-Performance Neural Networks for Class-Imbalanced Data
    Bansal, Arpit, Goldblum, Micah, Cherepanova, Valeriia, Schwarzschild, Avi, Bruss, C Bayan, and Goldstein, Tom
    arXiv preprint arXiv:2106.09643 2021

2020

    2019

    1. Pag-net: Progressive attention guided depth super-resolution network
      Bansal, Arpit, Jonna, Sankaraganesh, and Sahay, Rajiv R
      arXiv preprint arXiv:1911.09878 2019