The success of machine learning is fueled by the increasing availability of computing power
and large training datasets. The training data is used to learn new models or update existing …

Due to the greatly improved capabilities of devices, massive data, and increasing concern
about data privacy, Federated Learning (FL) has been increasingly considered for …

Multimodal contrastive pretraining has been used to train multimodal representation models,
such as CLIP, on large amounts of paired image-text data. However, previous studies have …

In a backdoor attack, an adversary inserts maliciously constructed backdoor examples into a
training set to make the resulting model vulnerable to manipulation. Defending against such …

A powerful category of (invisible) data poisoning attacks modify a subset of training
examples by small adversarial perturbations to change the prediction of certain test-time …

Given the availability of abundant data, deep learning models have been advanced and
become ubiquitous in the past decade. In practice, due to many different reasons (eg …

To mitigate recent insidious backdoor attacks on deep learning models, advances have
been made by the research community. Nonetheless, state-of-the-art defenses are either …

Backdoor attacks against deep learning have been widely studied. Various attack
techniques have been proposed for different domains and paradigms, eg, image, point …

We introduce camouflaged data poisoning attacks, a new attack vector that arises in the
context of machine unlearning and other settings when model retraining may be induced. An …

Recent advances in generative AI have significantly expanded into the realms of art and
music. This development has opened up a vast realm of possibilities, pushing the …