The world's first drone war between nuclear-armed neighbours has erupted in South Asia. On Thursday, India accused Pakistan of launching waves of drones and missiles at three military bases in Indian territory and Indian-administered Kashmir - an allegation Islamabad swiftly denied. Pakistan claimed it had shot down 25 Indian drones in recent hours. Experts say the tit-for-tat attacks mark a dangerous new phase in the decades-old rivalry, as both sides exchange not just artillery but unmanned weapons across a volatile border. As Washington and other global powers urge restraint, the region is teetering on the edge of escalation, with drones - silent, remote and deniable - opening a new chapter in the India-Pakistan conflict.

Tensions continue to mount as India and Pakistan traded accusations and attacks across their frontier in Kashmir overnight. New Delhi and Islamabad accused one another on Friday of launching drone attacks as well as "numerous ceasefire violations" over the Line of Control (LoC) in the disputed territory. The ongoing hostilities have provoked further calls for restraint as the risk of an escalation between the two nuclear powers grows. Pakistan launched "multiple attacks" using drones and other munitions along India's western border on Thursday night and early Friday, the Indian army said, claiming it had repelled the attacks and responded forcefully, although it did not provide details. Islamabad has denied any cross-border attacks and instead accused Indian forces of sending drones into Pakistani territory, killing at least two civilians.

In the days ahead of the proposed truce, Moscow and Kyiv exchanged a barrage of strikes. Flights at airports across Russia were cancelled and some 60,000 passengers left stranded in the wake of Ukrainian drone attacks. Heavy restrictions are in place in the centre of Moscow as Russia prepares to mark the Soviet Union's victory over Nazi Germany. Russia says 27 world leaders are attending the event, with thousands of troops marching on Red Square ahead of a parade of some of Russia's latest weaponry. Brazil's Luiz Inácio Lula da Silva and Venezuelan President Nicolas Maduro are among the assembled guests, along with Serbian President Aleksandar Vucic and Robert Fico, Slovakia's prime minister who is the only European Union leader to travel to Moscow. Ukraine's Volodymyr Zelensky had earlier warned that he could not guarantee the safety of anyone attending the event and has urged heads of state not to travel to Moscow.

The cybersecurity company that became a household name after causing a massive global IT outage last year has announced it will cut 5% of its workforce in part due to "AI efficiency". In a note to staff earlier this week, released in stock market filings in the US, CrowdStrike's chief executive, George Kurtz, announced that 500 positions, or 5% of its workforce, would be cut globally, citing AI efficiencies created in the business. "We're operating in a market and technology inflection point, with AI reshaping every industry, accelerating threats, and evolving customer needs," he said. Kurtz said AI "flattens our hiring curve, and helps us innovate from idea to product faster", adding it "drives efficiencies across both the front and back office". "AI is a force multiplier throughout the business," he said.

Expanding the application of large language models (LLMs) to societal life, instead of primary function only as auxiliary assistants to communicate with only one person at a time, necessitates LLMs' capabilities to independently play roles in multi-user, multi-turn social agent tasks within complex social settings. However, currently the capability has not been systematically measured with available benchmarks. To address this gap, we first introduce an agent task leveling framework grounded in sociological principles. Concurrently, we propose a novel benchmark, How Social Is It (we call it HSII below), designed to assess LLM's social capabilities in comprehensive social agents tasks and benchmark representative models. HSII comprises four stages: format parsing, target selection, target switching conversation, and stable conversation, which collectively evaluate the communication and task completion capabilities of LLMs within realistic social interaction scenarios dataset, HSII-Dataset. The dataset is derived step by step from news dataset. We perform an ablation study by doing clustering to the dataset. Additionally, we investigate the impact of chain of thought (COT) method on enhancing LLMs' social performance. Since COT cost more computation, we further introduce a new statistical metric, COT-complexity, to quantify the efficiency of certain LLMs with COTs for specific social tasks and strike a better trade-off between measurement of correctness and efficiency. Various results of our experiments demonstrate that our benchmark is well-suited for evaluating social skills in LLMs.

The medical device industry has significantly advanced by integrating sophisticated electronics like microchips and field-programmable gate arrays (FPGAs) to enhance the safety and usability of life-saving devices. These complex electro-mechanical systems, however, introduce challenging failure modes that are not easily detectable with conventional methods. Effective fault detection and mitigation become vital as reliance on such electronics grows. This paper explores three generative machine learning-based approaches for fault detection in medical devices, leveraging sensor data from surgical staplers,a class 2 medical device. Historically considered low-risk, these devices have recently been linked to an increasing number of injuries and fatalities. The study evaluates the performance and data requirements of these machine-learning approaches, highlighting their potential to enhance device safety.

Framing in media critically shapes public perception by selectively emphasizing some details while downplaying others. With the rise of large language models in automated news and content creation, there is growing concern that these systems may introduce or even amplify framing biases compared to human authors. In this paper, we explore how framing manifests in both out-of-the-box and fine-tuned LLM-generated news content. Our analysis reveals that, particularly in politically and socially sensitive contexts, LLMs tend to exhibit more pronounced framing than their human counterparts. In addition, we observe significant variation in framing tendencies across different model architectures, with some models displaying notably higher biases. These findings point to the need for effective post-training mitigation strategies and tighter evaluation frameworks to ensure that automated news content upholds the standards of balanced reporting.

Machine learning has become a powerful tool for enhancing data assimilation. While supervised learning remains the standard method, reinforcement learning (RL) offers unique advantages through its sequential decision-making framework, which naturally fits the iterative nature of data assimilation by dynamically balancing model forecasts with observations. We develop RL-DAUNCE, a new RL-based method that enhances data assimilation with physical constraints through three key aspects. First, RL-DAUNCE inherits the computational efficiency of machine learning while it uniquely structures its agents to mirror ensemble members in conventional data assimilation methods. Second, RL-DAUNCE emphasizes uncertainty quantification by advancing multiple ensemble members, moving beyond simple mean-state optimization. Third, RL-DAUNCE's ensemble-as-agents design facilitates the enforcement of physical constraints during the assimilation process, which is crucial to improving the state estimation and subsequent forecasting. A primal-dual optimization strategy is developed to enforce constraints, which dynamically penalizes the reward function to ensure constraint satisfaction throughout the learning process. Also, state variable bounds are respected by constraining the RL action space. Together, these features ensure physical consistency without sacrificing efficiency. RL-DAUNCE is applied to the Madden-Julian Oscillation, an intermittent atmospheric phenomenon characterized by strongly non-Gaussian features and multiple physical constraints. RL-DAUNCE outperforms the standard ensemble Kalman filter (EnKF), which fails catastrophically due to the violation of physical constraints. Notably, RL-DAUNCE matches the performance of constrained EnKF, particularly in recovering intermittent signals, capturing extreme events, and quantifying uncertainties, while requiring substantially less computational effort.

The rapid advancement of generative models, such as Stable Diffusion, raises a key question: how can synthetic data from these models enhance predictive modeling? While they can generate vast amounts of datasets, only a subset meaningfully improves performance. We propose a novel end-to-end framework that generates and systematically filters synthetic data through domain-specific statistical methods, selectively integrating high-quality samples for effective augmentation. Our experiments demonstrate consistent improvements in predictive performance across various settings, highlighting the potential of our framework while underscoring the inherent limitations of generative models for data augmentation. Despite the ability to produce large volumes of synthetic data, the proportion that effectively improves model performance is limited.

Rotational symmetry plays a central role in physics, providing an elegant framework to describe how the properties of 3D objects -- from atoms to the macroscopic scale -- transform under the action of rigid rotations. Equivariant models of 3D point clouds are able to approximate structure-property relations in a way that is fully consistent with the structure of the rotation group, by combining intermediate representations that are themselves spherical tensors. The symmetry constraints however make this approach computationally demanding and cumbersome to implement, which motivates increasingly popular unconstrained architectures that learn approximate symmetries as part of the training process. In this work, we explore a third route to tackle this learning problem, where equivariant functions are expressed as the product of a scalar function of the point cloud coordinates and a small basis of tensors with the appropriate symmetry. We also propose approximations of the general expressions that, while lacking universal approximation properties, are fast, simple to implement, and accurate in practical settings.