Designing effective embodied multi-agent systems is critical for solving complex real-world tasks across domains. Due to the complexity of multi-agent embodied systems, existing methods fail to automatically generate safe and efficient training data for such systems. To this end, we propose the concept of compositional constraints for embodied multi-agent systems, addressing the challenges arising from collaboration among embodied agents. We design various interfaces tailored to different types of constraints, enabling seamless interaction with the physical world. Leveraging compositional constraints and specifically designed interfaces, we develop an automated data collection framework for embodied multi-agent systems and introduce the first benchmark for embodied multi-agent manipulation, RoboFactory. Based on RoboFactory benchmark, we adapt and evaluate the method of imitation learning and analyzed its performance in different difficulty agent tasks. Furthermore, we explore the architectures and training strategies for multi-agent imitation learning, aiming to build safe and efficient embodied multi-agent systems.

Expressing confidence is challenging for embodied agents navigating dynamic multimodal environments, where uncertainty arises from both perception and decision-making processes. We present the first work investigating embodied confidence elicitation in open-ended multimodal environments. We introduce Elicitation Policies, which structure confidence assessment across inductive, deductive, and abductive reasoning, along with Execution Policies, which enhance confidence calibration through scenario reinterpretation, action sampling, and hypothetical reasoning. Evaluating agents in calibration and failure prediction tasks within the Minecraft environment, we show that structured reasoning approaches, such as Chain-of-Thoughts, improve confidence calibration. However, our findings also reveal persistent challenges in distinguishing uncertainty, particularly under abductive settings, underscoring the need for more sophisticated embodied confidence elicitation methods.

Long-horizon tasks in robotic manipulation present significant challenges in reinforcement learning (RL) due to the difficulty of designing dense reward functions and effectively exploring the expansive state-action space. However, despite a lack of dense rewards, these tasks often have a multi-stage structure, which can be leveraged to decompose the overall objective into manageable subgoals. In this work, we propose DEMO3, a framework that exploits this structure for efficient learning from visual inputs. Specifically, our approach incorporates multi-stage dense reward learning, a bi-phasic training scheme, and world model learning into a carefully designed demonstration-augmented RL framework that strongly mitigates the challenge of exploration in long-horizon tasks. Our evaluations demonstrate that our method improves data-efficiency by an average of 40% and by 70% on particularly difficult tasks compared to state-of-the-art approaches. We validate this across 16 sparse-reward tasks spanning four domains, including challenging humanoid visual control tasks using as few as five demonstrations.

Is it possible to expose the vulnerabilities of a given robot policy with respect to changes in environmental factors such as lighting, visual distractors, and object placement without performing hardware evaluations in these scenarios? As we seek to deploy robots in environments with ever-increasing complexity, it becomes imperative to develop scalable methods for predicting how well they will generalize when faced with unseen scenarios. Performing hardware evaluations to discover vulnerabilities -- which can depend in surprising ways on the specifics of policy training and architecture -- is often prohibitively expensive to set up and execute, especially when the goal is to test the limits of safe deployment in a sufficiently diverse set of scenarios. As an example, consider a visuomotor diffusion policy [1] trained to perform pick-and-place tasks via behavior cloning (Figure 1). The policy is trained with a large dataset: over 3K+ demonstrations with varied objects, locations, and visual distractors. Will the policy generalize well to a change in the height of the table by a few centimeters (as one may plausibly predict due to the variations in 2D object locations in the training dataset) compared to when a human is standing closer to the table than seen during training? If so, what is the absolute degradation of the success rate in each case? As it turns out, the above prediction is incorrect: the success rate of the policy degrades from 65% under nominal conditions to 10% by changing the table height, and remains roughly constant with a human close to the table. Predicting the relative and absolute impact of other factors (e.g., lighting, table backgrounds, object distractors; Figure 1) can be even more challenging.

Are the current credit underwriting policies not inclusive? They may not be, as these policies are drafted by human beings, like you and me, who are inherently biased and therefore prone to making rules that may discriminate against certain individuals or communities without the intention to do so. To address this issue effectively, the federal law in the US makes it illegal for a lender to deny credit or offer different terms based on protected traits like race, color, or religion. But do we have the same rules in India? As of today, we, unfortunately, don't have any such specific regulations in place.

Several policy options exist, or have been proposed, to further responsible artificial intelligence (AI) development and deployment. Institutions, including U.S. government agencies, states, professional societies, and private and public sector businesses, are well positioned to implement these policies. However, given limited resources, not all policies can or should be equally prioritized. We define and review nine suggested policies for furthering responsible AI, rank each policy on potential use and impact, and recommend prioritization relative to each institution type. We find that pre-deployment audits and assessments and post-deployment accountability are likely to have the highest impact but also the highest barriers to adoption. We recommend that U.S. government agencies and companies highly prioritize development of pre-deployment audits and assessments, while the U.S. national legislature should highly prioritize post-deployment accountability. We suggest that U.S. government agencies and professional societies should highly prioritize policies that support responsible AI research and that states should highly prioritize support of responsible AI education. We propose that companies can highly prioritize involving community stakeholders in development efforts and supporting diversity in AI development. We advise lower levels of prioritization across institutions for AI ethics statements and databases of AI technologies or incidents. We recognize that no one policy will lead to responsible AI and instead advocate for strategic policy implementation across institutions.

The vast majority of FDA-approved medical devices enabled by artificial intelligence or machine learning are concentrated in radiology and cardiovascular care, according to an analysis by Rock Health. Rock Health used data from FDA clearances and approvals from 1997 to 2021 to determine where these devices are used the most. Here are the AI/machine-learning enabled devices by therapeutic area, the Oct. 8 report found:

Paul Rosenzweig is the founder of Red Branch Consulting PLLC, a homeland security consulting company and a Senior Advisor to The Chertoff Group. Mr. Rosenzweig formerly served as Deputy Assistant Secretary for Policy in the Department of Homeland Security. He is a Professorial Lecturer in Law at George Washington University, a Senior Fellow in the Tech, Law & Security program at American University, and a Board Member of the Journal of National Security Law and Policy.

AI is a general-purpose technology that has the potential to improve the welfare and well-being of people, to contribute to positive sustainable global economic activity, to increase innovation and productivity, and to help respond to key global challenges. It is deployed in many sectors ranging from production, finance and transport to healthcare and security. Alongside benefits, AI also raises challenges for our societies and economies, notably regarding economic shifts and inequalities, competition, transitions in the labour market, and implications for democracy and human rights. The OECD has undertaken empirical and policy activities on AI in support of the policy debate over the past two years, starting with a Technology Foresight Forum on AI in 2016 and an international conference on AI: Intelligent Machines, Smart Policies in 2017. The Organisation also conducted analytical and measurement work that provides an overview of the AI technical landscape, maps economic and social impacts of AI technologies and their applications, identifies major policy considerations, and describes AI initiatives from governments and other stakeholders at national and international levels.

Artificial intelligence, machine learning, and advanced analytics are critical to the challenge of producing intelligence from the enormous amounts of data available to the Intelligence Community. This day-long symposium will focus on the challenge of "Big Data" and what the IC needs to do to address it. Panels will examine the IC's AIM Strategy; the importance of the Pentagon's newly created Chief Digital and AI Officer; challenges and success stories with operationalizing AI; and the capabilities needed to make it all work. Media Policy This event is open to the press. Contact [email protected] for more information.