Ping-yeh Chiang

江秉燁


Ph.D. student
Department of Computer Science
University of Maryland, College Park

Bio

I am a fifth year PhD student in Computer Science at University of Maryland, under the supervision of Prof. Tom Goldstein. My thesis focuses on ensuring the reliability and safety of ML systems under extreme scenarios.

News

  • [Jan 2023] Our paper on the implicity bias of neural network is accepted by the ICLR 2023.
  • [May 2022] Our paper on certified neural network watermark is accepted by ICML 2022.
  • [Mar 2022] Our paper studying reproducibility of neural network is accepted by CVPR 2022.
  • [Aug 2021] Our paper on data poisoning through adversarial example is accepted by NeurIPS 2021.
  • [Jan 2021] Our paper on hardware aware quantization is accepted by ICLR 2021.
  • [Aug 2020] Our paper on certified object detection through randomized smoothing is accepted by NeurIPS 2020.
  • [Aug 2020] Our paper on certifying strategyproof auction network is accepted by NeurIPS 2020.
  • [Jan 2020] Our paper on certified defense against patch attack is accepted by ICLR 2020.

    • Publications

  • Ping-yeh Chiang, Yipin Zhou, Omid Poursaeed, Satya Narayan Shukla, Tom Goldstein, Ser-Nam Lim, "Universal Pyramid Adversarial Training for Improved ViT Performance", Preprint.
  • Ping-yeh Chiang, Renkun Ni, David Yu Miller, Arpit Bansal, Jonas Geiping, Micah Goldblum, Tom Goldstein, "Loss Landscapes are All You Need: Neural Network Generalization Can Be Explained Without the Implicit Bias of Gradient Descent", in the proceeding of the International Conference on Learning Representations (ICLR), 2022. Spotlight [paper][code]
  • Gowthami Somepalli, Liam Fowl, Arpit Bansal, Ping Yeh-Chiang, Yehuda Dar, Richard Baraniuk, Micah Goldblum, Tom Goldstein, "Can neural nets learn the same model twice? Investigating reproducibility and double descent from the decision boundary perspective", in the proceeding of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2022. Oral [paper][code]
  • Arpit Amit Bansal*, Ping-yeh Chiang*, Michael Curry, Hossein Souri, Rama Chellappa, John P Dickerson, Rajiv Jain, Tom Goldstein, "Certified Watermarks for Neural Networks", in the proceeding of the International Conference on Machine Learning (ICML), 2022. Spotlight [paper]
  • Liam Fowl*, Micah Goldblum*, Ping-yeh Chiang*, Jonas Geiping, Wojtek Czaja, Tom Goldstein, "Adversarial Examples Make Strong Poisons", in the proceeding of the Conference on Neural Information Processing Systems (NeurIPS), 2021. [paper]
  • Ping-yeh Chiang, Michael J. Curry, Ahmed Abdelkader, Aounon Kumar, John Dickerson, Tom Goldstein, "Detection as Regression: Certified Object Detection by Median Smoothing", in the proceeding of the Conference on Neural Information Processing Systems (NeurIPS), 2020. [paper][code]
  • Michael J Curry*, Ping-Yeh Chiang*, Tom Goldstein, John Dickerson, "Certifying Strategyproof Auction Networks", in the proceeding of the Conference on Neural Information Processing Systems (NeurIPS), 2020. [paper]
  • Ping-Yeh Chiang, Jonas Geiping, Micah Goldblum, Tom Goldstein, Renkun Ni, Steven Reich, Ali Shafahi, "WITCHcraft: Efficient PGD attacks with random step size", in the proceeding of IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), 2020.[paper]
  • Ping-Yeh Chiang*, Renkun Ni*, Ahmed Abdelkader, Chen Zhu, Christoph Studor, Tom Goldstein, "Certified Defenses for Adversarial Patches", in the proceeding of the International Conference on Learning Representations (ICLR), 2020.[paper][code]

    • Contacts

    Iribe Center, Office 2116
    College Park, MD 20740
    pchiang-{AT}-umd-DOT-edu

    You'll Never Walk Alone


    Ping © last update April 2023