We introduce WildTeaming, an automatic LLM safety red-teaming framework that mines in-the-wild user-chatbot interactions to discover 5.7K unique clusters of novel jailbreak tactics, and then composes multiple tactics for systematic exploration of novel jailbreaks. Compared to prior work that performed red-teaming via recruited human workers, gradient-based optimization, or iterative revision with LLMs, our work investigates jailbreaks from chatbot users who were not specifically instructed to break the system. WildTeaming reveals previously unidentified vulnerabilities of frontier LLMs, resulting in up to 4.6x more diverse and successful adversarial attacks compared to state-of-the-art jailbreak methods. While many datasets exist for jailbreak evaluation, very few open-source datasets exist for jailbreak training, as safety training data has been closed even when model weights are open. With WildTeaming we create WildJailbreak, a large-scale open-source synthetic safety dataset with 262K vanilla (direct request) and adversarial (complex jailbreak) prompt-response pairs. To mitigate exaggerated safety behaviors, WildJailbreak provides two contrastive types of queries: 1) harmful queries (vanilla & adversarial) and 2) benign queries that resemble harmful queries in form but contain no harm. As WildJailbreak considerably upgrades the quality and scale of existing safety resources, it uniquely enables us to examine the scaling effects of data and the interplay of data properties and model capabilities during safety training. Through extensive experiments, we identify the training properties that enable an ideal balance of safety behaviors: appropriate safeguarding without over-refusal, effective handling of vanilla and adversarial queries, and minimal, if any, decrease in general capabilities. All components of WildJailbeak contribute to achieving balanced safety behaviors of models.

Location data is frequently collected from populations and shared in aggregate form to guide policy and decision making. However, the prevalence of aggregated data also raises the privacy concern of membership inference attacks (MIAs). MIAs infer whether an individual's data contributed to the aggregate release. Although effective MIAs have been developed for aggregate location data, these require access to an extensive auxiliary dataset of individual traces over the same locations, which are collected from a similar population. This assumption is often impractical given common privacy practices surrounding location data. To measure the risk of an MIA performed by a realistic adversary, we develop the first Zero Auxiliary Knowledge (ZK) MIA on aggregate location data, which eliminates the need for an auxiliary dataset of real individual traces. Instead, we develop a novel synthetic approach, such that suitable synthetic traces are generated from the released aggregate. We also develop methods to correct for bias and noise, to show that our synthetic-based attack is still applicable when privacy mechanisms are applied prior to release. Using two large-scale location datasets, we demonstrate that our ZK MIA matches the state-of-the-art Knock-Knock (KK) MIA across a wide range of settings, including popular implementations of differential privacy (DP) and suppression of small counts. Furthermore, we show that ZK MIA remains highly effective even when the adversary only knows a small fraction (10%) of their target's location history. This demonstrates that effective MIAs can be performed by realistic adversaries, highlighting the need for strong DP protection.

Robustness remains a paramount concern in deep reinforcement learning (DRL), with randomized smoothing emerging as a key technique for enhancing this attribute. However, a notable gap exists in the performance of current smoothed DRL agents, often characterized by significantly low clean rewards and weak robustness. In response to this challenge, our study introduces innovative algorithms aimed at training effective smoothed robust DRL agents. We propose S-DQN and S-PPO, novel approaches that demonstrate remarkable improvements in clean rewards, empirical robustness, and robustness guarantee across standard RL benchmarks. Notably, our S-DQN and S-PPO agents not only significantly outperform existing smoothed agents by an average factor of $2.16\times$ under the strongest attack, but also surpass previous robustly-trained agents by an average factor of $2.13\times$. This represents a significant leap forward in the field. Furthermore, we introduce Smoothed Attack, which is $1.89\times$ more effective in decreasing the rewards of smoothed agents than existing adversarial attacks.

The topic of fairness in AI, as debated in the FATE (Fairness, Accountability, Transparency, and Ethics in AI) communities, has sparked meaningful discussions in the past years. However, from a legal perspective, particularly from the perspective of European Union law, many open questions remain. Whereas algorithmic fairness aims to mitigate structural inequalities at design-level, European non-discrimination law is tailored to individual cases of discrimination after an AI model has been deployed. The AI Act might present a tremendous step towards bridging these two approaches by shifting non-discrimination responsibilities into the design stage of AI models. Based on an integrative reading of the AI Act, we comment on legal as well as technical enforcement problems and propose practical implications on bias detection and bias correction in order to specify and comply with specific technical requirements.

In the rapidly evolving landscape of artificial intelligence (AI), the collaboration between human intelligence and AI systems, known as Human-AI (HAI) Teaming, has emerged as a cornerstone for advancing problem-solving and decision-making processes. The advent of Large Pre-trained Models (LPtM) has significantly transformed this landscape, offering unprecedented capabilities by leveraging vast amounts of data to understand and predict complex patterns. This paper surveys the pivotal integration of LPtMs with HAI, emphasizing how these models enhance collaborative intelligence beyond traditional approaches. It examines the potential of LPtMs in augmenting human capabilities, discussing this collaboration for AI model improvements, effective teaming, ethical considerations, and their broad applied implications in various sectors. Through this exploration, the study sheds light on the transformative impact of LPtM-enhanced HAI Teaming, providing insights for future research, policy development, and strategic implementations aimed at harnessing the full potential of this collaboration for research and societal benefit.

Large Language Models (LLMs) have shown exceptional results on current benchmarks when working individually. The advancement in their capabilities, along with a reduction in parameter size and inference times, has facilitated the use of these models as agents, enabling interactions among multiple models to execute complex tasks. Such collaborations offer several advantages, including the use of specialized models (e.g. coding), improved confidence through multiple computations, and enhanced divergent thinking, leading to more diverse outputs. Thus, the collaborative use of language models is expected to grow significantly in the coming years. In this work, we evaluate the behavior of a network of models collaborating through debate under the influence of an adversary. We introduce pertinent metrics to assess the adversary's effectiveness, focusing on system accuracy and model agreement. Our findings highlight the importance of a model's persuasive ability in influencing others. Additionally, we explore inference-time methods to generate more compelling arguments and evaluate the potential of prompt-based mitigation as a defensive strategy.

Being able to carry out complicated vision language reasoning tasks in 3D space represents a significant milestone in developing household robots and human-centered embodied AI. In this work, we demonstrate that a critical and distinct challenge in 3D vision language reasoning is situational awareness, which incorporates two key components: (1) The autonomous agent grounds its self-location based on a language prompt. (2) The agent answers open-ended questions from the perspective of its calculated position. To address this challenge, we introduce SIG3D, an end-to-end Situation-Grounded model for 3D vision language reasoning. We tokenize the 3D scene into sparse voxel representation and propose a language-grounded situation estimator, followed by a situated question answering module. Experiments on the SQA3D and ScanQA datasets show that SIG3D outperforms state-of-the-art models in situation estimation and question answering by a large margin (e.g., an enhancement of over 30% on situation estimation accuracy). Subsequent analysis corroborates our architectural design choices, explores the distinct functions of visual and textual tokens, and highlights the importance of situational awareness in the domain of 3D question answering.

Recent advances in multimodal imaging acquisition techniques have allowed us to measure different aspects of brain structure and function. Multimodal fusion, such as linked independent component analysis (LICA), is popularly used to integrate complementary information. However, it has suffered from missing data, commonly occurring in neuroimaging data. Therefore, in this paper, we propose a Full Information LICA algorithm (FI-LICA) to handle the missing data problem during multimodal fusion under the LICA framework. Built upon complete cases, our method employs the principle of full information and utilizes all available information to recover the missing latent information. Our simulation experiments showed the ideal performance of FI-LICA compared to current practices. Further, we applied FI-LICA to multimodal data from the Alzheimer's Disease Neuroimaging Initiative (ADNI) study, showcasing better performance in classifying current diagnosis and in predicting the AD transition of participants with mild cognitive impairment (MCI), thereby highlighting the practical utility of our proposed method.

It's hard not to laugh at NASA's blooper reel of astronauts falling and bouncing in slow motion on the moon. But coping with inertia where gravity is one-sixth that of Earth is no laughing matter when you're wearing a constricting space suit and need to finish an exhausting task. So mechanical engineering professor Harry Asada (center) and colleagues are developing wearable robotic limbs to help astronauts get back on their feet after a fall. Based on the "SuperLimbs" Asada designed to assist construction workers and ship builders, the limbs extend from a backpack that would also contain the astronaut's life support system along with a controller and motors to provide power. As part of NASA's planned Artemis mission, astronauts will be expected to build the first permanent moon base--a physically demanding project with a high risk of falls during multiple extended extravehicular activities (EVAs).

In early 2023, following an international conference that included dialogue with China, the United States released a "Political Declaration on Responsible Military Use of Artificial Intelligence and Autonomy," urging states to adopt sensible policies that include ensuring ultimate human control over nuclear weapons. Yet the notion of "human control" itself is hazier than it might seem. If humans authorized a future AI system to "stop an incoming nuclear attack," how much discretion should it have over how to do so? The challenge is that an AI general enough to successfully thwart such an attack could also be used for offensive purposes. We need to recognize the fact that AI technologies are inherently dual-use.