Deep neural networks (DNNs) have shown unprecedented success in object
d...
Backdoor learning has become an emerging research area towards building ...
Recent studies show that prompt tuning can better leverage the power of ...
In this work, besides improving prediction accuracy, we study whether
pe...
Pre-trained language models allowed us to process downstream tasks with ...
Efficient and automated design of optimizers plays a crucial role in
ful...
Federated learning (FL) has recently attracted increasing attention from...
Despite the efficiency and scalability of machine learning systems, rece...
Predictor-based algorithms have achieved remarkable performance in the N...
Differentiable Neural Architecture Search is one of the most popular Neu...
Large-batch training has become a commonly used technique when training
...
Enhancing model robustness under new and even adversarial environments i...
Developing robust models against adversarial perturbations has been an a...
This paper proposes a novel differentiable architecture search method by...
Adversarial training has become one of the most effective methods for
im...
We propose an algorithm to enhance certified robustness of a deep model
...
We study the most practical problem setup for evaluating adversarial
rob...
This work proposes a novel algorithm to generate natural language advers...
We study the problem of attacking a machine learning model in the hard-l...
Derivative-free optimization has become an important technique used in
m...
Crafting adversarial examples has become an important technique to evalu...
Recent studies have revealed the vulnerability of deep neural networks -...