Recent years have seen a growing concern about climate change and its impacts. While Earth System Models (ESMs) can be invaluable tools for studying the impacts of climate change, the complex coupling processes encoded in ESMs and the large amounts of data produced by these models, together with the high internal variability of the Earth system, can obscure important source-to-impact relationships. This paper presents a novel and efficient unsupervised data-driven approach for detecting statistically-significant impacts and tracing spatio-temporal source-impact pathways in the climate through a unique combination of ideas from anomaly detection, clustering and Natural Language Processing (NLP). Using as an exemplar the 1991 eruption of Mount Pinatubo in the Philippines, we demonstrate that the proposed approach is capable of detecting known post-eruption impacts/events. We additionally describe a methodology for extracting meaningful sequences of post-eruption impacts/events by using NLP to efficiently mine frequent multivariate cluster evolutions, which can be used to confirm or discover the chain of physical processes between a climate source and its impact(s).

Crowdsourced annotations of data play a substantial role in the development of Artificial Intelligence (AI). It is broadly recognised that annotations of text data can contain annotator bias, where systematic disagreement in annotations can be traced back to differences in the annotators' backgrounds. Being unaware of such annotator bias can lead to representational bias against minority group perspectives and therefore several methods have been proposed for recognising bias or preserving perspectives. These methods typically require either a substantial number of annotators or annotations per data instance. In this study, we propose a simple method for handling bias in annotations without requirements on the number of annotators or instances. Instead, we ask annotators about their beliefs of other annotators' judgements of an instance, under the hypothesis that these beliefs may provide more representative and less biased labels than judgements. The method was examined in two controlled, survey-based experiments involving Democrats and Republicans (n=1,590) asked to judge statements as arguments and then report beliefs about others' judgements. The results indicate that bias, defined as systematic differences between the two groups of annotators, is consistently reduced when asking for beliefs instead of judgements. Our proposed method therefore has the potential to reduce the risk of annotator bias, thereby improving the generalisability of AI systems and preventing harm to unrepresented socio-demographic groups, and we highlight the need for further studies of this potential in other tasks and downstream applications.

With the rapid development of black-box machine learning (ML) models, such as deep neural networks or gradient boosting trees, the need for explanations of their decisions has emerged. This demand has been driven by the increasing implementation of opaque models, in high-risk and critical areas like medicine, healthcare, industry, and law, which laid the foundation for modern research on explainable and interpretable artificial intelligence (XAI). Scientists' efforts in designing XAI algorithms have been further supported by political initiatives such as DARPA's XAI challenge [1], the European Union's GDPR [2], and more recently, the EU AI Act [3]. The shared goal of all these initiatives is to improve the transparency of AI systems, thereby promoting their adoption in areas where trust in AI is not fully established or where the transparency of decisions is crucial for legal and safety reasons. However, as XAI algorithms have been advanced, a new discussion has been initiated, addressing the fundamental challenge of ensuring that the explanations generated by these algorithms are comprehensible to humans. This triggered research on the evaluation of XAI [4], drawing attention from social sciences, which argued that much of the effort in XAI relies solely on researchers' intuition about what constitutes a good explanation. They emphasized that human factors should be integral to the design and evaluation of XAI to ensure its reliability [5]. Recognizing individual human abilities to comprehend algorithmically generated explanations is crucial, as these abilities can vary significantly based on personal information competencies. Additionally, there is a lack of established multidisciplinary methods for measuring these capabilities, as well as datasets that facilitate reproducible evaluations or comprehensive analyses.

Ukraine has struck a manufacturer of military explosives deep inside Russian territory overnight, as well as storage infrastructure at a military airfield in the Lipetsk region, Kyiv's General Staff has said in a statement. For their part, Russian air defence units downed 110 Ukrainian drones over the country, Russia's Ministry of Defence said Sunday, including one over the Moscow region, 43 over the border region of Kursk, and 27 over the southwestern Lipetsk region. The Russian SHOT Telegram channel reported that drones attempted to strike the Ya. The explosives plant, one of the largest manufacturers of its kind used by Russian forces in the war that Moscow launched against Ukraine in February 2022, is subject to sanctions by the United States and the European Union. Such large-scale aerial attacks are still relatively rare on Russia. Kyiv's General Staff said in a post on Telegram the Sverdlov factory had been making chemical components for artillery ammunition and aerial bombs, adding that it was still assessing the damage from its attack.

After losing her husband, "Beatrice" turned to an online dating site for seniors during the COVID-19 pandemic. She quickly matched with and fell hard for a person she thought was a 66-year-old Spanish lumberjack who looked uncannily like her husband. "I was missing not having him here to talk about, you know, what was going on in the world and everything," Beatrice, who asked that her real name not be used, told Homeland Security Investigations (HSI). "So, somebody suggested to go online through a dating service… and this guy's pictures show up and he's just, you know, no George Clooney, nothing gorgeous, but in fact, he had a resemblance to my husband." The man spent about four months texting and calling the woman before he felt he had gained her trust – then, he began asking her to wire him money.

Ukraine launched a series of drones targeting Moscow and western Russia, according to regional officials. Russian air defence units downed 110 Ukrainian drones over Russia, the Ministry of Defence said, including one over the Moscow region, 43 over the border region of Kursk, and 27 over the southwestern Lipetsk region. Russia's air defence units destroyed at least one drone flying towards the capital, Moscow Mayor Sergei Sobyanin said on the Telegram messaging app, while drone debris sparked several short-lived fires in Lipetsk, the regional governor said on the app. No injuries or significant damage were reported from the attacks. Four firefighters suffered minor shrapnel wounds in a Ukrainian drone attack in an industrial zone in the city of Dzerzhinsk in Russia's Nizhny Novgorod region, the regional governor said.

Black box attacks, where adversaries have limited knowledge of the target model, pose a significant threat to machine learning systems. Adversarial examples generated with a substitute model often suffer from limited transferability to the target model. While recent work explores ranking perturbations for improved success rates, these methods see only modest gains. We propose a novel strategy called PEAS that can boost the transferability of existing black box attacks. PEAS leverages the insight that samples which are perceptually equivalent exhibit significant variability in their adversarial transferability. Our approach first generates a set of images from an initial sample via subtle augmentations. We then evaluate the transferability of adversarial perturbations on these images using a set of substitute models. Finally, the most transferable adversarial example is selected and used for the attack. Our experiments show that PEAS can double the performance of existing attacks, achieving a 2.5x improvement in attack success rates on average over current ranking methods. We thoroughly evaluate PEAS on ImageNet and CIFAR-10, analyze hyperparameter impacts, and provide an ablation study to isolate each component's importance.

As large language models (LLMs) continue to evolve, their potential use in automating cyberattacks becomes increasingly likely. With capabilities such as reconnaissance, exploitation, and command execution, LLMs could soon become integral to autonomous cyber agents, capable of launching highly sophisticated attacks. In this paper, we introduce novel defense strategies that exploit the inherent vulnerabilities of attacking LLMs. By targeting weaknesses such as biases, trust in input, memory limitations, and their tunnel-vision approach to problem-solving, we develop techniques to mislead, delay, or neutralize these autonomous agents. We evaluate our defenses under black-box conditions, starting with single prompt-response scenarios and progressing to real-world tests using custom-built CTF machines. Our results show defense success rates of up to 90\%, demonstrating the effectiveness of turning LLM vulnerabilities into defensive strategies against LLM-driven cyber threats.

We propose WHoW, an evaluation framework for analyzing the facilitation strategies of moderators across different domains/scenarios by examining their motives (Why), dialogue acts (How) and target speaker (Who). Using this framework, we annotated 5,657 moderation sentences with human judges and 15,494 sentences with GPT-4o from two domains: TV debates and radio panel discussions. Comparative analysis demonstrates the framework's cross-domain generalisability and reveals distinct moderation strategies: debate moderators emphasise coordination and facilitate interaction through questions and instructions, while panel discussion moderators prioritize information provision and actively participate in discussions. Our analytical framework works for different moderation scenarios, enhances our understanding of moderation behaviour through automatic large-scale analysis, and facilitates the development of moderator agents.

Teleoperation can transfer human perception and cognition to a slave robot to cope with some complex tasks, in which the agility and flexibility of the interface play an important role in mapping human intention to the robot. In this paper, we developed an agile large-workspace teleoperation interface by estimating human arm behavior. Using the wearable sensor, namely the inertial measurement unit and surface electromyography armband, we can capture the human arm motion and force information, thereby intuitively controlling the manipulation of the robot. The control principle of our wearable interface includes two parts: (1) the arm incremental kinematics and (2) the grasping recognition. Moreover, we developed a teleoperation framework with a time synchronization mechanism for the real-time application. We conducted experimental comparisons with a versatile haptic device (Omega 7) to verify the effectiveness of our interface and framework. Seven subjects are invited to complete three different tasks: free motion, handover, and pick-and-place action (each task ten times), and the total number of tests is 420. Objectively, we used the task completion time and success rate to compare the performance of the two interfaces quantitatively. In addition, to quantify the operator experience, we used the NASA Task Load Index to assess their subjective feelings. The results showed that the proposed interface achieved a competitive performance with a better operating experience.