Adversarial attacks on graphs have posed a major threat to the robustness of graph machine learning (GML) models. Naturally, there is an ever-escalating arms race between attackers and defenders. However, the strategies behind both sides are often not fairly compared under the same and realistic conditions. To bridge this gap, we present the Graph Robustness Benchmark (GRB) with the goal of providing a scalable, unified, modular, and reproducible evaluation for the adversarial robustness of GML models. GRB standardizes the process of attacks and defenses by 1) developing scalable and diverse datasets, 2) modularizing the attack and defense implementations, and 3) unifying the evaluation protocol in refined scenarios. By leveraging the GRB pipeline, the end-users can focus on the development of robust GML models with automated data processing and experimental evaluations. To support open and reproducible research on graph adversarial learning, GRB also hosts public leaderboards across different scenarios. As a starting point, we conduct extensive experiments to benchmark baseline techniques. GRB is open-source and welcomes contributions from the community.

In the last decade, there have been significant advances in multi-agent reinforcement learning (MARL) but there are still numerous challenges, such as high sample complexity and slow convergence to stable policies, that need to be overcome before wide-spread deployment is possible. However, many real-world environments already, in practice, deploy sub-optimal or heuristic approaches for generating policies. An interesting question which arises is how to best use such approaches as advisors to help improve reinforcement learning in multi-agent domains. In this paper, we provide a principled framework for incorporating action recommendations from online sub-optimal advisors in multi-agent settings. We describe the problem of ADvising Multiple Intelligent Reinforcement Agents (ADMIRAL) in nonrestrictive general-sum stochastic game environments and present two novel Q-learning based algorithms: ADMIRAL - Decision Making (ADMIRAL-DM) and ADMIRAL - Advisor Evaluation (ADMIRAL-AE), which allow us to improve learning by appropriately incorporating advice from an advisor (ADMIRAL-DM), and evaluate the effectiveness of an advisor (ADMIRAL-AE). We analyze the algorithms theoretically and provide fixed-point guarantees regarding their learning in general-sum stochastic games. Furthermore, extensive experiments illustrate that these algorithms: can be used in a variety of environments, have performances that compare favourably to other related baselines, can scale to large state-action spaces, and are robust to poor advice from advisors.

Drones are apparently turning into assassination tools. According to CBS News and Reuters, Iraqi Prime Minister Mustafa al-Kadhimi says he survived a drone-based assassination attempt today (November 7th) at his home in Baghdad's highly secure Green Zone. The country's Interior Ministry said the attack involved three drones, including at least one bomb-laden vehicle. Six bodyguards were injured during the incident, and an official speaking talking to Reuters claimed security forces obtained the remnants of a small drone at the scene. While the Iraqi government publicly said it was "premature" to identify culprits, CBS sources suspected the perpetrators belonged to pro-Iranian militias that have used similar tactics against Erbil International Airport and the US Embassy.

America's national security organizations have begun applying AI to more quickly and effectively produce intelligence assessments. Speaking at the GovernmentCIO Media & Research AI: National Security virtual event, Director of the National Security Agency (NSA) Research Directorate Mark Segal discussed how these new capacities are assisting intelligence analysts in better processing and sorting large quantities of often complex and disparate information. In outlining the NSA's research priorities, Segal noted that both AI and machine-learning capacities already showed promise for better organizing the large pools of variable data their analysts sort through in producing regular assessments. "One of the challenges that we have found AI to be particularly useful for is looking through the sheer amount of data that's created every day on this planet. Our analysts are looking at some of this data trying to understand it, and understand what its implications are for national security. The amount of data that we have to sort is going up pretty dramatically, but the number of people that we have who are actually looking at this data is pretty constant. So we're constantly looking for tools and technologies to help our analysts more effectively go through huge piles of data," Segal said.

On Friday, the radical transparency group DDoSecrets released hundreds of hours of police helicopter surveillance footage. It's unclear who originally obtained the data, or what that person's motivations were, but the trove shows how extensive law enforcement's eye-in-the-sky has become, and how high-fidelity its cameras are. Privacy advocates also say the incident underscores that authorities don't do nearly enough to protect sensitive data, and have retention policies that are far too lax. In other aerial news: For the first time, intelligence officials say, a consumer drone likely attempted to disrupt the US power grid. The July 2020 incident took place at a power substation in Pennsylvania; a DJI Mavic 2 quadcopter outfitted with nylon ropes and copper wire seemed determined to cause a short circuit, but crash-landed on a nearby roof before it reached its apparent target.

Iraqi Prime Minister Mustafa al-Kadhemi escaped unhurt from an "assassination attempt" in which an explosives-packed drone hit his Baghdad residence early Sunday, a new escalation in the country's post-election turmoil. Washington condemned the "apparent act of terrorism" while Iraqi President Barham Saleh called the attack, which was not immediately claimed by any group, an attempted "coup against the constitutional system". Kadhemi, aged 54 and in power since May 2020, appealed for "calm and restraint" before chairing a meeting at his office in the high-security Baghdad Green Zone, where the overnight attack took place. Three drones were launched from near a Tigris River bridge but two were intercepted, according to security sources, who said two bodyguards were wounded. Gunfire rang out and smoke rose from the Green Zone after the strike, which the premier's office labelled a "failed assassination attempt".

Its permanent searchable database is accessed by more than 2,400 police agencies including US Immigration and Customs Enforcement (ICE). Clearview AI uses an algorithm to extract unique features in the human face to create a trackable "faceprint." The EU has stringent personal privacy standards, including the GDPR and the Right to Be Forgotten, which are in conflict with Clearview AI's methods. Facial recognition technology has received substantial backlash for its racial bias and inaccuracy, which have resulted in numerous false arrests. At least 14 US cities have banned facial recognition use, and Maine and Massachusetts passed statewide laws banning the tech from law enforcement.

Iraq's Prime Minister Mustafa al-Kadhemi said he was unhurt and appealed for "calm and restraint" after a drone attack on his residence early Sunday heightened political tensions in the war-scarred country. The attack in Baghdad's Green Zone was the first to target the residence of Kadhemi, who has been in power since May 2020. It came as Iraq's political parties negotiate alliances over who will run the next government after elections last month. That vote saw the Conquest (Fatah) Alliance, the political arm of the pro-Iran Hashed al-Shaabi paramilitary network, suffer a substantial decline in its parliamentary seats, leading the group to denounce the outcome as "fraud". The big winner, with more than 70 seats according to the initial count, was the movement of Moqtada Sadr, a Shiite Muslim preacher who campaigned as a nationalist and critic of Iran.

BAGHDAD – Iraqi Prime Minister Mustafa al-Kadhimi escaped unharmed in an assassination attempt by armed drone in Baghdad on Sunday, officials said, in an incident that dramatically raises tension in the country weeks after a general election disputed by Iran-backed militia groups. Six members of Kadhimi's personal protection force stationed outside his residence in the Green Zone were wounded, security sources said. Three drones were used in the attack, including two that were intercepted and downed by security forces while a third drone hit the residence, state news agency INA quoted a spokesman for the interior ministry as saying. A spokesman for the armed forces commander in chief said the security situation was stable inside the fortified Green Zone, which houses the residence, government buildings and foreign embassies, after the drone attack. No group immediately claimed responsibility for the attack.

Teams across NASA and the Department of Labor are accelerating artificial intelligence and data capabilities to prepare for high-performance computing and supercharge existing tools and technologies, agency leaders said during GovernmentCIO Media & Research's AI Gov: National Security virtual event last week. Krista Kinnard, Labor's chief of emerging technologies, said high-performance computing will help the agency accelerate forms processing. It also wants to boost natural language processing and data protection by leveraging lessons learned from agencies like NASA and the Defense Department. "What we've learned from these agencies around protecting privacy, protecting data, ensuring that data cannot be hacked, or leaked or altered in any way, is incredibly relevant to how we meet our mission," Kinnard said. As agencies prepare for the "compute" stage, data serves as a foundational element and must be consumable by a model and stored in an accessible location, Kinnard said.