How is Russia replenishing its military? What is a'coalition of the willing'? How China forgot promises and'debts' to Ukraine How are Europe, the US pulling apart on Ukraine? Anti-aircraft units downed seven Ukrainian drones headed for Moscow early on Friday, according to the Russian capital's mayor Sergei Sobyanin. Russian forces have taken control of the settlement of Sosnivka in Ukraine's Dnipropetrovsk region, Russia's Defence Ministry said on Thursday.

The FBI has released new footage showing the suspect in the killing of US activist Charlie Kirk running across a roof - from where the fatal shot was fired - before dropping to the ground and crossing a road. As the authorities continue their search, BBC Verify's Nick Beake has been looking at the footage and what we know about his escape. Watch: Key moments from RFK Jr's heated Senate hearing The US health secretary faced questions on Covid deaths and vaccines a week after firing the head of the Centers for Disease Control and Prevention. The group said they are making their own list of Jeffrey Epstein's associates and called for the release of all files related to the investigation. The molten magma streamed down one of the world's most active volcanoes and put on another fiery show on the Island of Hawaii.

'Looks like a toy, but it's real': BBC examines a downed Russian drone At least three Russian drones were shot down in Poland's airspace during attacks on Ukraine, the Polish prime minister said on Wednesday. The BBC's Sarah Rainsford has been looking at the exact type of Russian drone that flew into Poland, and is proving a massive challenge for Ukraine's territorial defence forces. The BBC's Sarah Rainsford says Sunday's attack caused a huge amount of damage. One of Kyiv's main government buildings was hit in overnight missile and drone strikes by Russia. 'The hit was very hard': Eyewitness in second carriage shares video of crash moment The incident in Lisbon's funicular has left 16 dead and multiple injured.

Utah authorities have released CCTV footage showing the Charlie Kirk shooting suspect fleeing the scene at Utah Valley University. The video shows a figure dressed in black running across a roof before dropping down to the ground and fleeing towards a wooded area. Authorities say the suspect was wearing Converse shoes, sunglasses and a distinctive black T-shirt with an American flag and an eagle. Watch: Key moments from RFK Jr's heated Senate hearing The US health secretary faced questions on Covid deaths and vaccines a week after firing the head of the Centers for Disease Control and Prevention. The group said they are making their own list of Jeffrey Epstein's associates and called for the release of all files related to the investigation.

In-context learning (ICL) allows a language model to improve its problem-solving capability when provided with suitable information in context. Since the choice of in-context information can be determined based on the problem itself, in-context learning is analogous to human learning from teachers in a classroom. Recent works (Didolkar et al., 2024a; 2024b) show that ICL performance can be improved by leveraging a frontier large language model's (LLM) ability to predict required skills to solve a problem, popularly referred to as an LLM's metacognition, and using the recommended skills to construct necessary in-context examples. While this skill-based strategy boosts ICL performance in larger models, its gains on small language models (SLMs) have been minimal, highlighting a performance gap in ICL capabilities. We investigate this gap and show that skill-based prompting can hurt SLM performance on easy questions by introducing unnecessary information, akin to cognitive overload. To address this, we introduce AdaptMI, an adaptive approach to selecting skill-based in-context Math Instructions for SLMs. Inspired by cognitive load theory from human pedagogy, our method only introduces skill-based examples when the model performs poorly. We further propose AdaptMI+, which adds examples targeted to the specific skills missing from the model's responses. On 5-shot evaluations across popular math benchmarks and five SLMs (1B--7B; Qwen, Llama), AdaptMI+ improves accuracy by up to 6% over naive skill-based strategies.

Retrieval-Augmented Generation (RAG) systems in the legal domain face a critical challenge: standard, flat-text retrieval is blind to the hierarchical, diachronic, and causal structure of law, leading to anachronistic and unreliable answers. This paper introduces the Structure-Aware Temporal Graph RAG (SAT-Graph RAG), an ontology-driven framework designed to overcome these limitations by explicitly modeling the formal structure and diachronic nature of legal norms. We ground our knowledge graph in a formal, LRMoo-inspired model that distinguishes abstract legal Works from their versioned Expressions. We model temporal states as efficient aggregations that reuse the versioned expressions (CTVs) of unchanged components, and we reify legislative events as first-class Action nodes to make causality explicit and queryable. This structured backbone enables a unified, planner-guided query strategy that applies explicit policies to deterministically resolve complex requests for (i) point-in-time retrieval, (ii) hierarchical impact analysis, and (iii) auditable provenance reconstruction. Through a case study on the Brazilian Constitution, we demonstrate how this approach provides a verifiable, temporally-correct substrate for LLMs, enabling higher-order analytical capabilities while drastically reducing the risk of factual errors. The result is a practical framework for building more trustworthy and explainable legal AI systems.

Large Language Models (LLMs) exhibit significant performance variations depending on the linguistic and cultural context in which they are applied. This disparity signals the necessity of mature evaluation frameworks that can assess their capabilities in specific regional settings. In the case of Portuguese, existing evaluations remain limited, often relying on translated datasets that may not fully capture linguistic nuances or cultural references. Meanwhile, native Portuguese-language datasets predominantly focus on structured national exams or sentiment analysis of social media interactions, leaving gaps in evaluating broader linguistic understanding. To address this limitation, we introduce BRoverbs, a dataset specifically designed to assess LLM performance through Brazilian proverbs. Proverbs serve as a rich linguistic resource, encapsulating cultural wisdom, figurative expressions, and complex syntactic structures that challenge the model comprehension of regional expressions. BRoverbs aims to provide a new evaluation tool for Portuguese-language LLMs, contributing to advancing regionally informed benchmarking. The benchmark is available at https://huggingface.co/datasets/Tropic-AI/BRoverbs.

Solutions to the Algorithm Selection Problem (ASP) in machine learning face the challenge of high computational costs associated with evaluating various algorithms' performances on a given dataset. To mitigate this cost, the meta-learning field can leverage previously executed experiments shared in online repositories such as OpenML. OpenML provides an extensive collection of machine learning experiments. However, an analysis of OpenML's records reveals limitations. It lacks diversity in pipelines, specifically when exploring data preprocessing steps/blocks, such as scaling or imputation, resulting in limited representation. Its experiments are often focused on a few popular techniques within each pipeline block, leading to an imbalanced sample. To overcome the observed limitations of OpenML, we propose PIPES, a collection of experiments involving multiple pipelines designed to represent all combinations of the selected sets of techniques, aiming at diversity and completeness. PIPES stores the results of experiments performed applying 9,408 pipelines to 300 datasets. It includes detailed information on the pipeline blocks, training and testing times, predictions, performances, and the eventual error messages. This comprehensive collection of results allows researchers to perform analyses across diverse and representative pipelines and datasets. PIPES also offers potential for expansion, as additional data and experiments can be incorporated to support the meta-learning community further. The data, code, supplementary material, and all experiments can be found at https://github.com/cynthiamaia/PIPES.git.

We propose a data-driven method for approximating real-valued functions on smooth manifolds, building on the Diffusion Maps framework under the manifold hypothesis. Given pointwise evaluations of a function, the method constructs a smooth extension to the ambient space by exploiting diffusion geometry and its connection to the heat equation and the Laplace-Beltrami operator. To address the computational challenges of high-dimensional data, we introduce a dimensionality reduction strategy based on the low-rank structure of the distance matrix, revealed via singular value decomposition (SVD). In addition, we develop an online updating mechanism that enables efficient incorporation of new data, thereby improving scalability and reducing computational cost. Numerical experiments, including applications to sparse CT reconstruction, demonstrate that the proposed methodology outperforms classical feedforward neural networks and interpolation methods in terms of both accuracy and efficiency.

Don't worry too much about planet-warming emissions, the US Energy Secretary has told the BBC, because within five years AI will have enabled the harnessing of nuclear fusion - the energy that powers the sun and stars. Chris Wright told me in an interview that he expected the technology to deliver power to electricity grids around the world within eight to 15 years and that it would rapidly become a big driver of greenhouse gas reductions. His claims will likely surprise even enthusiasts for the technology. Harnessing the energy released when atoms fuse together could produce vast amounts of low carbon energy but most scientists believe commercial fusion power plants are still a long way off. With artificial intelligence and what's going on at the national labs and private companies in the United States, we will have that approach about how to harness fusion energy multiple ways within the next five years, said Mr Wright.