T
Tinashe Handina
Publications - 3
Citations - 16
Tinashe Handina is an academic researcher. The author has contributed to research in topics: Proxy (statistics) & Robustness (computer science). The author has an hindex of 1, co-authored 2 publications receiving 6 citations.
Papers
More filters
Proceedings ArticleDOI
Chasing Convex Bodies and Functions with Black-Box Advice
TL;DR: Two novel algorithms are proposed that bypass the common paradigm of algorithms that switch between the decisions of the advice and a competitive algorithm by exploiting the problem’s convexity.
Posted Content
Improving Adversarial Robustness Using Proxy Distributions.
Vikash Sehwag,Saeed Mahloujifar,Tinashe Handina,Sihui Dai,Chong Xiang,Mung Chiang,Prateek Mittal +6 more
TL;DR: This article showed that the difference between the adversarial robustness of a classifier on the proxy and original training dataset distribution is upper bounded by the conditional Wasserstein distance between them.
Posted Content
Robust Learning Meets Generative Models: Can Proxy Distributions Improve Adversarial Robustness?.
Vikash Sehwag,Saeed Mahloujifar,Tinashe Handina,Sihui Dai,Chong Xiang,Mung Chiang,Prateek Mittal +6 more
TL;DR: The authors proposed a robust discrimination approach, which measures the distinguishability of synthetic and real samples under adversarial perturbations, and proposed a set of 10 million most beneficial synthetic samples for robust training on the CIFAR-10 dataset.