A brave eight-year-old boy helped save his grandad after the pair drifted more than two miles (3km) from the coast on a capsized kayak. Marley and his granscha, David Dai Jones, from Mountain Ash, Rhondda Cynon Taf, had been kayaking off Fontygary in the Vale of Glamorgan on 27 May when they capsized and were unable to get back onboard. Dai managed to help Marley back onto the kayak but could not climb back on himself. He remained in the water holding on as the pair drifted in the strong Bristol Channel currents. Despite the frightening situation, Marley remained calm and used a mobile phone kept in a waterproof pouch to contact his nan on shore, who called 999.

Recent incidents heighten anxieties that hybrid warfare tactics could trigger military confrontation with Russia. Lithuanian armed special forces and members of the Lithuanian Riflemen's Union take part in a military exercise in central Lithuania [File: Nils Adler/Al Jazeera] A member of the Lithuanian Riflemen's Union joins in military exercises in central Lithuania [File: Nils Adler/Al Jazeera] Along the forests and marshlands that separate the Baltic states from Russia and Belarus, workers are digging anti-tank ditches, pouring concrete bunkers and erecting rows of dragon's teeth - jagged concrete obstacles designed to slow and channel advancing armour - to buy precious time in the event of an attack. Russia's full-scale invasion of Ukraine in 2022 reignited old fears in Estonia, Latvia and Lithuania, where memories of Soviet rule remain close to the surface. In the years since, those fears have been channelled into preparation. Defence budgets have surged, military exercises have intensified, and new fortifications have emerged even as daily life largely continues as normal.

Ukrainian drones have struck a historic museum in Russia-annexed Sevastopol in Crimea, igniting a roof fire, as Russian authorities slashed nighttime train schedules amid intensifying air attacks across the peninsula and deep into Russia. Sevastopol's Russian-installed governor, Mikhail Razvozhayev, announced the damage on Telegram early on Wednesday. "This building is not just a museum, it is a symbol of resilience, which has repeatedly taken the blows of the enemy." Razvozhayev said that during World War II's Siege of Sevastopol, "the Panorama building was subjected to massed bombing by German aviation". He declared: "The enemy will pay for this sacrilege!"

Uncrewed ground vehicles have already been tested for defending the front line by the Ukrainian military. There's a received piece of wisdom among militaries around the world that whatever new technologies appear, in the end, foot soldiers are what matters. As British Army officer Field Marshal Archibald Wavell put it shortly after the second world war: "All battles and all wars are won in the end by the infantryman." This may now finally be changing. Robots in battle are about to reach a critical point for Ukraine. In May, it began the mass production of Legit, a low-cost robot capable of carrying a machine gun.

Ni, who moved to the UK in 2019 to study, was targeted by what she believes is a pro-regime bot. Ni, who moved to the UK in 2019 to study, was targeted by what she believes is a pro-regime bot. A high-profile Chinese activist in the UK who was inundated with deepfake posts on X portraying her as a sexually promiscuous drug addict was told that the abuse did not breach the rules of Elon Musk's platform. Apple Peiqing Ni, the 27-year-old founder of the UK-based China Dissent Network, had been advised by UK police to complain to the US-headquartered platform after she was targeted by what she believes is a pro-regime bot. The abuse included 12 posts tagging Ni and containing fake photographs and videos of her.

Conformal Prediction (CP) provides robust uncertainty guarantees for predictive models, but is typically applied post hoc, which misaligns model training with the conformal goal of producing efficient (i.e, narrow) intervals. We propose SPACR (Single-Pass Adaptive Conformal Regressor), a novel method for directly training uncertainty-aware regressors within a differentiable loss. SPACR jointly optimizes efficiency and validity without batch-splitting or a predefined confidence levels during training. As a result, a single SPACR model yields valid prediction intervals at multiple confidence levels during inference, avoiding the costly retraining required by methods like DOICR. Experiments on diverse datasets show that SPACR consistently gives tighter intervals and better coverage-efficiency trade-offs compared to standard CP and DOICR, while significantly reducing computational costs.

Existing methods rely predominantly on token-level signals, leaving the geometric structure of intermediate hidden states underused. In this paper, we take the geometric complexity of hidden-state matrices as a measure of the global uncertainty of LLMs, while treating token-level uncertainty estimation as a local metric. We show that hidden-state geometric entropy (global uncertainty) and token-level entropy (local uncertainty) are statistically near-orthogonal, capturing distinct failure regimes for reliability prediction. In particular, global geometry recovers the confident-but-wrong failure mode that local signals systematically miss. Building on this, we propose Global-Local Uncertainty (GLU), an unsupervised, single-pass score that fuses the two signals via a multiplicative gate. Across three model families and six benchmarks, GLU matches or outperforms all unsupervised baselines while requiring only a single forward pass and remaining length-normalized and architecture-agnostic. Code is available on https://github.com/qcri/GLU.git.

Reconstruction-based methods are widely used for time series anomaly detection, where models are trained to reconstruct subsequences, and anomalies are identified through reconstruction errors. However, reported results are often hard to compare due to heterogeneous evaluation practices and underspecified inference procedures. In this paper, we revisit reconstruction-based anomaly detection in the univariate offline setting and study the role of the inference stride, which controls whether subsequences are processed as disjoint windows or with overlap. We propose a unified training, tuning, and multi-seed evaluation protocol on the curated TSB-AD benchmark, and study how overlapping inference affects anomaly detection performance for a range of reconstruction models, including PCA-based baselines, DLinear, an AutoEncoder, TimesNet, and Transformer variants. The results show that across all models, overlapping windows yield consistent improvements, with average relative gain up to +28%, and can alter method rankings. We further analyze variability across datasets, random seeds, and hyperparameter configurations. Finally, we complement the benchmark study with an evaluation on the full UCR archive using localization criteria aligned with sliding-window reconstruction. Overall, our results highlight that reconstruction-based anomaly detection performance depends not only on model architecture and training, but also on inference choices, motivating a clear and reproducible protocol. Our results show that reconstructionbased baselines achieve strong performance on both TSB-AD and UCR benchmarks, supporting them as competitive and practical approaches for univariate time series anomaly detection.

We explore whether intrinsic symmetries of the training data lead to conserved quantities during gradient-flow training of neural networks. Under the assumption that the loss function is analytic and non-polynomial, we prove that data symmetries generically do not induce any additional integrals of motion. For mean squared error (MSE) loss, on the other hand, there are situations in which data augmentation yields extra conserved quantities. We build a framework, utilizing tensorizable networks to describe this phenomenon. Tensorizable networks are a family of architectures whose dependence on parameters and inputs can be separated using an intermediate representation. They include linear and Figure 1: A display of how data symmetry can give polynomial networks, as well as Lightning At-rise to conservation laws. The top row shows the tention.

Mixture-of-Experts large language models (LLMs) scale efficiently through sparse activation, yet their deployment is fundamentally constrained by the large static parameter footprint of experts. Existing compression approaches either remove entire experts, disrupting routing topology and harming performance, or rely on unstructured weight pruning with limited practical efficiency. To address the limitations, we propose TENP, a structured Trapezoidal ExpertNeuron Pruning framework. Using a few samples, we identify and retain important experts, while applying expert neuron pruning (ENP) to less important experts, reserving model parameters in a trapezoidal pattern from shallow to deep layers. When evaluating expert importance, we jointly consider both the magnitude of the expert output and its ability to change the direction of the input vector. For ENP, we measure each neuron's projected contribution to the expert output to identify and retain important neurons. We conduct extensive experiments on the Qwen and DeepSeek models. Under a routing expert sparsity of 40% and an average of 63.76% activated expert parameters, the DeepSeek model suffers only a 1-point drop in accuracy compared to the full-parameter model. Moreover, it outperforms the full-parameter model by 10% on code generation tasks.