From pilotless jets engaging in dogfights to huge undersea vessels ferrying troops, the Pentagon is pushing to increase the U.S. military's use of automation. Defense moves are outpacing commercial automation efforts in the air, on the ground and beneath the waves as officials seek to counter American adversaries' technological advances, according to current and former national-security and industry officials. That progress--highlighted in cockpits managed primarily by computers, totally autonomous helicopters and automated aerial-refueling tankers--is likely to show up in future civilian aircraft, advanced air-traffic-control systems and a range of drone applications. Skeptics worry automated systems sometimes reflect software developers' desire to incorporate new capabilities without full testing. They point to examples of high-profile stumbles ranging from glitch-prone radio communication systems to software problems that have deprived pilots of adequate oxygen at the controls of jet fighters.

Modeling geophysical systems as dynamical systems and regressing their vector field from data is a simple way to learn emulators for such systems. We show that when the kernel of these emulators is also learned from data (using kernel flows, a variant of cross-validation), then the resulting data-driven models are not only faster than equation-based models but are easier to train than neural networks such as the long short-term memory neural network. In addition, they are also more accurate and predictive than the latter. When trained on observational data for the global sea-surface temperature, considerable gains are observed by the proposed technique in comparison to classical partial differential equation-based models in terms of forecast computational cost and accuracy. When trained on publicly available re-analysis data for temperatures in the North-American continent, we see significant improvements over climatology and persistence based forecast techniques.

On a cold December night in 1950, red-baiting Sen. Joe McCarthy spent a charity dinner at Washington's Sulgrave Club trading insults with liberal journalist Drew Pearson. McCarthy had attacked Pearson on the floor of the Senate, calling for a boycott of his radio show. Pearson had attacked McCarthy on air and in his newspaper column, accusing the senator of lying about communist infiltration of the American government. McCarthy had recklessly accused the State Department of harboring hundreds of communists, sparking a massive investigation and an ongoing purge. After dinner, the two ran into each other in the cloakroom and their conflict turned physical.

Reinforcement Learning (RL) algorithms have led to recent successes in solving complex games, such as Atari or Starcraft, and to a huge impact in real-world applications, such as cybersecurity or autonomous driving. In the side of the drawbacks, recent works have shown how the performance of RL algorithms decreases under the influence of soft changes in the reward function. However, little work has been done about how sensitive these disturbances are depending on the aggressiveness of the attack and the learning exploration strategy. In this paper, we propose to fill this gap in the literature analyzing the effects of different attack strategies based on reward perturbations, and studying the effect in the learner depending on its exploration strategy. In order to explain all the behaviors, we choose a sub-class of MDPs: episodic, stochastic goal-only-rewards MDPs, and in particular, an intelligible grid domain as a benchmark. In this domain, we demonstrate that smoothly crafting adversarial rewards are able to mislead the learner, and that using low exploration probability values, the policy learned is more robust to corrupt rewards. Finally, in the proposed learning scenario, a counterintuitive result arises: attacking at each learning episode is the lowest cost attack strategy.

Quantiphi, an applied AI and data science software and services company, today announced that it has been named an IDC Innovator in the IDC Innovators: Artificial Intelligence Services, 2020 ( Doc # US45733220, December 2020) report. Quantiphi is one of just four companies featured in the report, which covers a selection of vendors that offer an innovative new technology or a groundbreaking business model, or both in artificial intelligence (AI) services. "AI has quickly evolved from a'nice-to-have technology' to a business imperative, driving enterprise demand for expertise from solution design through production at scale," said Jennifer Hamel, Research Manager for IDC's Worldwide Services team. "Quantiphi approaches the AI services market in distinct ways, partnering with its clients to apply complex AI techniques to solve real business problems." The report acknowledges Quantiphi's broad portfolio of repeatable IP and accelerators, and strong partnerships with major AI technology providers (e.g., Google, AWS, and NVIDIA) to assemble and scale AI solutions for clients in a variety of industries, leveraging a talent pool of industry analysts, cloud/data engineers, and ML engineers.

As machine learning algorithms have been widely deployed across applications, many concerns have been raised over the fairness of their predictions, especially in high stakes settings (such as facial recognition and medical imaging). To respond to these concerns, the community has proposed and formalized various notions of fairness as well as methods for rectifying unfair behavior. While fairness constraints have been studied extensively for classical models, the effectiveness of methods for imposing fairness on deep neural networks is unclear. In this paper, we observe that these large models overfit to fairness objectives, and produce a range of unintended and undesirable consequences. We conduct our experiments on both facial recognition and automated medical diagnosis datasets using state-of-the-art architectures.

Academic timetabling is the task of scheduling courses to specific times in such a way that there are no conflicts. Most of the models considered in the literature assume that this conflict information is already known. However in many real life timetabling problems, courses are taught in multiple sections and until a student is assigned to a specific section of a course, the conflict information is not known. M.W. Carter [Car00] sums it up nicely "When courses are offered in multiple sections as they are at Waterloo, it creates a timetabling paradox. Students request a course, but timetabling assigns days and times to course sections. We cannot assign times to sections until we know which students are in each section. But we cannot assign students to sections until we know when the sections are timetabled!"

Semantic web technologies have shown their effectiveness, especially when it comes to knowledge representation, reasoning, and data integrations. However, the original semantic web vision, whereby machine readable web data could be automatically actioned upon by intelligent software web agents, has yet to be realised. In order to better understand the existing technological challenges and opportunities, in this paper we examine the status quo in terms of intelligent software web agents, guided by research with respect to requirements and architectural components, coming from that agents community. We start by collating and summarising requirements and core architectural components relating to intelligent software agent. Following on from this, we use the identified requirements to both further elaborate on the semantic web agent motivating use case scenario, and to summarise different perspectives on the requirements when it comes to semantic web agent literature. Finally, we propose a hybrid semantic web agent architecture, discuss the role played by existing semantic web standards, and point to existing work in the broader semantic web community any beyond that could help us to make the semantic web agent vision a reality.

For AI technology to fulfill its full promises, we must design effective mechanisms into the AI systems to support responsible AI behavior and curtail potential irresponsible use, e.g. in areas of privacy protection, human autonomy, robustness, and prevention of biases and discrimination in automated decision making. In this paper, we present a framework that provides computational facilities for parties in a social ecosystem to produce the desired responsible AI behaviors. To achieve this goal, we analyze AI systems at the architecture level and propose two decentralized cryptographic mechanisms for an AI system architecture: (1) using Autonomous Identity to empower human users, and (2) automating rules and adopting conventions within social institutions. We then propose a decentralized approach and outline the key concepts and mechanisms based on Decentralized Identifier (DID) and Verifiable Credentials (VC) for a general-purpose computational infrastructure to realize these mechanisms. We argue the case that a decentralized approach is the most promising path towards Responsible AI from both the computer science and social science perspectives.

This article is part of the Free Speech Project, a collaboration between Future Tense and the Tech, Law, & Security Program at American University Washington College of Law that examines the ways technology is influencing how we think about speech. Last summer's anti–police brutality protests represented the largest mass demonstration effort in American history. Since then, law enforcement departments nationwide have faced intense scrutiny for how they policed these historic protests. The repeated, egregious instances of violence against journalists and protesters are well documented and have driven widespread calls for systematic reform. These calls have focused in part on surveillance, after the police used sophisticated social media data monitoring, commandeered non-city camera networks, and tried other intrusive methods to identify suspects.