For a long time, researchers have sought artificial intelligence (AI) that matches or exceeds
human intelligence. AI agents, which are artificial entities capable of sensing the …

Adversarial training is one of the most effective approaches defending against adversarial
examples for deep learning models. Unlike other defense strategies, adversarial training …

Cross-entropy is a widely used loss function in applications. It coincides with the logistic loss
applied to the outputs of a neural network, when the softmax is used. But, what guarantees …

Pre-trained vision-language models (eg, CLIP) have shown promising zero-shot
generalization in many downstream tasks with properly designed text prompts. Instead of …

This work identifies 18 foundational challenges in assuring the alignment and safety of large
language models (LLMs). These challenges are organized into three different categories …

Recent work argues that robust training requires substantially larger datasets than those
required for standard classification. On CIFAR-10 and CIFAR-100, this translates into a …

Despite efforts to align large language models (LLMs) with human values, widely-used
LLMs such as GPT, Llama, Claude, and PaLM are susceptible to jailbreaking attacks …

As a research community, we are still lacking a systematic understanding of the progress on
adversarial robustness which often makes it hard to identify the most promising ideas in …

Understanding to what extent neural networks memorize training data is an intriguing
question with practical and theoretical implications. In this paper we show that in some …

Medical artificial intelligence (AI) systems have been remarkably successful, even
outperforming human performance at certain tasks. There is no doubt that AI is important to …