In this article, we propose a privacy-preserving image classification method that uses
encrypted images and an isotropic network, such as the vision transformer. The proposed …

Existing collaborative self-supervised learning (SSL) schemes are not suitable for cross-
client applications because of their expensive computation and large local data …

Privacy-preserving instance encoding aims to encode raw data into feature vectors without
revealing their privacy-sensitive information. When designed properly, these encodings can …

We initiate a formal study on the concept of learnable obfuscation and aim to answer the
following question: is there a type of data encoding that maintains the" learnability" of …

With the development of cloud services and machine learning, there has been an inevitable
need to enhance privacy and security when serving video recognition models. Although …

In this paper, we propose a privacy-preserving image classification method using block-wise
scrambled images and a modified ConvMixer. Conventional block-wise scrambled …

NeuraCrypt (Yara et al. arxiv 2021) is an algorithm that converts a sensitive dataset to an
encoded dataset so that (1) it is still possible to train machine learning models on the …

Many online communications systems use perceptual hash matching systems to detect illicit
files in user content. These systems employ specialized perceptual hash functions such as …

With the progression in sensor technology and the low-cost electronic circuits' proliferation,
the Internet of Things (IoT) is emerging as a promising technology for enhancing people life …

A large number of data are uploaded to cloud-based machine learning service providers. A
major concern is that such data often contain a lot more sensitive information than the …