South Korea sent drones across the border into North Korea for the first time Monday, an unprecedented tit-for-tat military move after Kim Jong Un's regime dispatched five unmanned aerial vehicles into its airspace. The exchange of drones, which briefly stopped flights from taking off at major airports near Seoul, came as Kim opened a major political meeting to set security, economic and political policy for the coming year, the official Korean Central News Agency reported Tuesday. He has spent the past year improving his atomic arsenal, showing no interest in returning to nuclear disarmament talks that have been stalled for three years. Kim's regime sent five drones across the border on Monday, the first time he has done so in more than five years. The first one crossed the border at 10:25 a.m. and returned after flying for about three hours.

How does Twitter use artificial intelligence and machine learning? Twitter uses large-scale machine learning and AI for sentiment analysis, bot analysis and detection of fake accounts, image classification and more. From Amazon to Instagram, Sephora, Microsoft, and Twitter, AI will shape the future of speech in America and beyond. The big question is not if they use it, but how it is being used, and what impact will this have on consumer privacy in the future. For the past fifteen years, I have been a national commentator on the politics of big tech and social media platforms. Social Media content decisions have become highly political, and artificial intelligence has proliferated this process at scale. But somewhere along the way, the public was left in the dark on just how large of a role machine learning plays in large-scale content operations in Silicon Valley. While the national conversation on free speech focuses on high-profile executives of tech companies and how content ...

President & CEO of BBB National Programs, a non-profit organization dedicated to fostering a more accountable, trustworthy marketplace. In recent years, government scrutiny over the use of artificial intelligence (AI) tools in the recruiting and hiring process has risen. Since I wrote about this topic last year, there has been significant activity within several federal government agencies regarding the use of AI and machine learning in the employment context. A better understanding of these actions can help business leaders reduce their risk of legal liability and better understand how to use AI and machine learning responsibly in their organizations. The Equal Employment Opportunity Commission (EEOC) has been particularly active through its EEOC initiative on AI and algorithmic fairness and its joint HIRE initiative with the U.S. Department of Labor.

Artificial intelligence (AI) is playing a massive role in cyber attacks and is proving both a "double-edged sword" and a "huge challenge," according to NATO. "Artificial intelligence allows defenders to scan networks more automatically, and fend off attacks rather than doing it manually. But the other way around, of course, it's the same game," David van Weel, NATO's Assistant Secretary-General for Emerging Security Challenges, told reporters earlier this month. Cyber attacks, both on national infrastructures and private companies, have ramped up exponentially and become a focal point since the war in Ukraine. NATO said this year that a cyber attack on any of its member states could trigger Article 5, meaning an attack on one member is considered an attack on all of them and could trigger a collective response.

Scientists and philosophers have debated whether humans can trust advanced artificial intelligence (AI) agents to respect humanity's best interests. Yet what about the reverse? Will advanced AI agents trust humans? Gauging an AI agent's trust in humans is challenging because--absent costs for dishonesty--such agents might respond falsely about their trust in humans. Here we present a method for incentivizing machine decisions without altering an AI agent's underlying algorithms or goal orientation. In two separate experiments, we then employ this method in hundreds of trust games between an AI agent (a Large Language Model (LLM) from OpenAI) and a human experimenter (author TJ). In our first experiment, we find that the AI agent decides to trust humans at higher rates when facing actual incentives than when making hypothetical decisions. Our second experiment replicates and extends these findings by automating game play and by homogenizing question wording. We again observe higher rates of trust when the AI agent faces real incentives. Across both experiments, the AI agent's trust decisions appear unrelated to the magnitude of stakes. Furthermore, to address the possibility that the AI agent's trust decisions reflect a preference for uncertainty, the experiments include two conditions that present the AI agent with a non-social decision task that provides the opportunity to choose a certain or uncertain option; in those conditions, the AI agent consistently chooses the certain option. Our experiments suggest that one of the most advanced AI language models to date alters its social behavior in response to incentives and displays behavior consistent with trust toward a human interlocutor when incentivized.

Many problems involve the use of models which learn probability distributions or incorporate randomness in some way. In such problems, because computing the true expected gradient may be intractable, a gradient estimator is used to update the model parameters. When the model parameters directly affect a probability distribution, the gradient estimator will involve score function terms. This paper studies baselines, a variance reduction technique for score functions. Motivated primarily by reinforcement learning, we derive for the first time an expression for the optimal state-dependent baseline, the baseline which results in a gradient estimator with minimum variance. Although we show that there exist examples where the optimal baseline may be arbitrarily better than a value function baseline, we find that the value function baseline usually performs similarly to an optimal baseline in terms of variance reduction. Moreover, the value function can also be used for bootstrapping estimators of the return, leading to additional variance reduction. Our results give new insight and justification for why value function baselines and the generalized advantage estimator (GAE) work well in practice.

Artificial intelligence (AI) in its various forms finds more and more its way into complex distributed systems. For instance, it is used locally, as part of a sensor system, on the edge for low-latency high-performance inference, or in the cloud, e.g. for data mining. Modern complex systems, such as connected vehicles, are often part of an Internet of Things (IoT). To manage complexity, architectures are described with architecture frameworks, which are composed of a number of architectural views connected through correspondence rules. Despite some attempts, the definition of a mathematical foundation for architecture frameworks that are suitable for the development of distributed AI systems still requires investigation and study. In this paper, we propose to extend the state of the art on architecture framework by providing a mathematical model for system architectures, which is scalable and supports co-evolution of different aspects for example of an AI system. Based on Design Science Research, this study starts by identifying the challenges with architectural frameworks. Then, we derive from the identified challenges four rules and we formulate them by exploiting concepts from category theory. We show how compositional thinking can provide rules for the creation and management of architectural frameworks for complex systems, for example distributed systems with AI. The aim of the paper is not to provide viewpoints or architecture models specific to AI systems, but instead to provide guidelines based on a mathematical formulation on how a consistent framework can be built up with existing, or newly created, viewpoints. To put in practice and test the approach, the identified and formulated rules are applied to derive an architectural framework for the EU Horizon 2020 project ``Very efficient deep learning in the IoT" (VEDLIoT) in the form of a case study.

Object detection models commonly deployed on uncrewed aerial systems (UAS) focus on identifying objects in the visible spectrum using Red-Green-Blue (RGB) imagery. However, there is growing interest in fusing RGB with thermal long wave infrared (LWIR) images to increase the performance of object detection machine learning (ML) models. Currently LWIR ML models have received less research attention, especially for both ground- and air-based platforms, leading to a lack of baseline performance metrics evaluating LWIR, RGB and LWIR-RGB fused object detection models. Therefore, this research contributes such quantitative metrics to the literature. The results found that the ground-based blended RGB-LWIR model exhibited superior performance compared to the RGB or LWIR approaches, achieving a mAP of 98.4%. Additionally, the blended RGB-LWIR model was also the only object detection model to work in both day and night conditions, providing superior operational capabilities. This research additionally contributes a novel labelled training dataset of 12,600 images for RGB, LWIR, and RGB-LWIR fused imagery, collected from ground-based and air-based platforms, enabling further multispectral machine-driven object detection research.

Anomaly detection on time series data is increasingly common across various industrial domains that monitor metrics in order to prevent potential accidents and economic losses. However, a scarcity of labeled data and ambiguous definitions of anomalies can complicate these efforts. Recent unsupervised machine learning methods have made remarkable progress in tackling this problem using either single-timestamp predictions or time series reconstructions. While traditionally considered separately, these methods are not mutually exclusive and can offer complementary perspectives on anomaly detection. This paper first highlights the successes and limitations of prediction-based and reconstruction-based methods with visualized time series signals and anomaly scores. We then propose AER (Auto-encoder with Regression), a joint model that combines a vanilla auto-encoder and an LSTM regressor to incorporate the successes and address the limitations of each method. Our model can produce bi-directional predictions while simultaneously reconstructing the original time series by optimizing a joint objective function. Furthermore, we propose several ways of combining the prediction and reconstruction errors through a series of ablation studies. Finally, we compare the performance of the AER architecture against two prediction-based methods and three reconstruction-based methods on 12 well-known univariate time series datasets from NASA, Yahoo, Numenta, and UCR. The results show that AER has the highest averaged F1 score across all datasets (a 23.5% improvement compared to ARIMA) while retaining a runtime similar to its vanilla auto-encoder and regressor components. Our model is available in Orion, an open-source benchmarking tool for time series anomaly detection.

Identifying cybersecurity discussions in open forums at scale is a topic of great interest for the purpose of mitigating and understanding modern cyber threats [1-3]. The challenge is that these discussions are typically quite noisy (i.e., they contain community known synonyms or acronyms or slang) and it is difficult to get labelled data in order to train resilient NLP (natural language processing) topic classifiers. Additionally, it is important that a tool that detects cybersecurity discussions in internet text sources is scalable and offers low errors rates (in particular, both low false negative rates and low false positive rates). In order to address the challenges of finding relevant cybersecurity labelled data, we use a technique that gathers posts or articles from different internet sources that have user defined topic labels. We then collect and label the training text as being cybersecurity related or not based on the subset of labels that the text source offers.