Event Type: Seminar
Location: Online - Zoom meeting
Time zone: AEST (GMT +11)
Who can attend: Everyone
Registration: At this link
Host Organisation: Monash University
Contact Details: muhammed.esgin@monash.edu
Details:
Machine learning has become increasingly prominent and is widely used in various applications in practice. Despite its great success, the integrity of machine learning predictions and accuracy is a rising concern. The reproducibility of machine learning models that are claimed to achieve high accuracy remains challenging, and the correctness and consistency of machine learning predictions in real products lack any security guarantees.
In this talk, Dr. Zhang will introduce some of his recent results on applying the cryptographic primitive of zero knowledge proofs to the domain of machine learning to address these issues. The protocols allow the owner of a machine learning model to convince others that the model computes a particular prediction on a data sample, or achieves a high accuracy on public datasets, without leaking any information about the machine learning model itself. He will talk about efficient zero knowledge proof protocols based on interactive proofs, and their applications on zero knowledge convolutional neural networks and decision trees.
Based on:
Jiaheng Zhang, Tianyi Liu, Weijie Wang, Yinuo Zhang, Dawn Song, Xiang Xie and Yupeng Zhang, 'Doubly Efficient Interactive Proofs for General Arithmetic Circuits with Linear Prover Time', eprint 2020/1247
Jiaheng Zhang, Zhiyong Fang, Yupeng Zhang and Dawn Song, 'Zero Knowledge Proofs for Decision Tree Predictions and Accuracy', ACM Conference on Computer and Communications Security (CCS) , 2020.
Tianyi Liu, Xiang Xie and Yupeng Zhang 'zkCNN: Zero Knowledge Proofs for Convolutional Neural Network Predictions and Accuracy', eprint 2021/673
About the Speaker
Yupeng Zhang
Assistant Professor, Texas A&M University
Yupeng Zhang is an Assistant Professor in the Computer Science and Engineering Department at Texas A&M University. His research is focused on zero knowledge proofs, secure multiparty computations, and their applications on privacy-preserving machine learning and zero knowledge machine learning. His work has been published at top security conferences including CCS, S&P, USENIX and CRYPTO. He is the recipient of Google PhD Fellowship, ACM SIGSAC Doctoral Dissertation Award Runners-up, and Distinguished Dissertation Award of ECE, University of Maryland. Before joining Texas A&M, he was a postdoctoral researcher at UC Berkeley hosted by Prof. Dawn Song. He obtained his PhD from University of Maryland and his advisors are Prof. Charalampos Papamanthou and Prof. Jonathan Katz.