Spanish Prime Minister Pedro Sánchez's decision to call a snap election for 29 November came on the back of a parliamentary defeat. On Friday, the raft of measures he had presented to congress aimed at improving conditions for tenants and trying to address the problem of soaring rentals was narrowly rejected, as parties on the right voted against them. That parliamentary result confirmed that Sánchez's left-wing coalition no longer commanded a majority. But as has so often been the case throughout Sánchez's career, the calling of the election was not a knee-jerk reaction but a carefully thought-out manouevre. His government has been struggling in recent months. His Socialist Party has been mired in corruption, with cases implicating former senior figures in the party, Santos Cerdán and José Luis Ábalos, while the former Socialist Prime Minister, José Luis Rodríguez Zapatero, is facing accusations of having benefited financially from the government's bailout of an airline.

Image caption, Sánchez's announcement comes after a series of measures to tackle the housing crisis was defeated in parliament on Friday Spanish Prime Minister Pedro Sánchez has called an early election for 29 November after recent defeats on legislation aimed at addressing the country's housing crisis. Speculation about a snap election had been building over the weekend after parliament rejected two decree laws put forward by the minority government to protect tenants in the rental market. The eviction of an 87-year-old woman from her home after she was unable to afford a sharp rent rise has sparked nationwide protests and increased pressure for housing reform. In a televised address on Monday, Sánchez said that to push reforms through he needed a broader progressive majority that can overcome the resistance of the elites. Housing has not been the only issue causing problems for Sánchez's leadership in recent months, with his party having been caught up in numerous corruption scandals and facing pressure on other cost-of-living issues and immigration.

Spain's Prime Minister Pedro Sanchez gives a speech during the World Governments Summit in Dubai on February 3, 2026. Spain's Prime Minister Pedro Sanchez gives a speech during the World Governments Summit in Dubai on February 3, 2026. The Spanish government has called for an investigation into social media giants X, Meta, and TikTok over their alleged role in producing and spreading AI-generated child sexual abuse material. The Council of Ministers will invoke Article 8 of the Organic Statute of the Public Prosecution Service to request that it investigate the crimes that X, Meta and TikTok may be committing through the creation and dissemination of child pornography by means of their AI, Spanish Prime Minister Pedro Sánchez wrote on X on Tuesday. Sánchez accused the platforms of "attacking the mental health, dignity and rights of our sons and daughters," saying that "the impunity of the giants must end."

Sparse Autoencoders (SAEs) can efficiently identify candidate monosemantic features from pretrained neural networks for galaxy morphology. We demonstrate this on Euclid Q1 images using both supervised (Zoobot) and new self-supervised (MAE) models. Our publicly released MAE achieves superhuman image reconstruction performance. While a Principal Component Analysis (PCA) on the supervised model primarily identifies features already aligned with the Galaxy Zoo decision tree, SAEs can identify interpretable features outside of this framework. SAE features also show stronger alignment than PCA with Galaxy Zoo labels. Although challenges in interpretability remain, SAEs provide a powerful engine for discovering astrophysical phenomena beyond the confines of human-defined classification.

We study the concentration of the Neural Tangent Kernel (NTK) $K_\theta : \mathbb{R}^{m_0} \times \mathbb{R}^{m_0} \to \mathbb{R}^{m_l \times m_l}$ of $l$-layer Multilayer Perceptrons (MLPs) $N : \mathbb{R}^{m_0} \times \Theta \to \mathbb{R}^{m_l}$ equipped with activation functions $\phi(s) = a s + b \vert s \vert$ for some $a,b \in \mathbb{R}$ with the parameter $\theta \in \Theta$ being initialized at the Edge Of Chaos (EOC). Without relying on the gradient independence assumption that has only been shown to hold asymptotically in the infinitely wide limit, we prove that an approximate version of gradient independence holds at finite width. Showing that the NTK entries $K_\theta(x_{i_1},x_{i_2})$ for $i_1,i_2 \in [1:n]$ over a dataset $\{x_1,\cdots,x_n\} \subset \mathbb{R}^{m_0}$ concentrate simultaneously via maximal inequalities, we prove that the NTK matrix $K(\theta) = [\frac{1}{n} K_\theta(x_{i_1},x_{i_2}) : i_1,i_2 \in [1:n]] \in \mathbb{R}^{nm_l \times nm_l}$ concentrates around its infinitely wide limit $\overset{\scriptscriptstyle\infty}{K} \in \mathbb{R}^{nm_l \times nm_l}$ without the need for linear overparameterization. Our results imply that in order to accurately approximate the limit, hidden layer widths have to grow quadratically as $m_k = k^2 m$ for some $m \in \mathbb{N}+1$ for sufficient concentration. For such MLPs, we obtain the concentration bound $\mathbb{P}( \Vert K(\theta) - \overset{\scriptscriptstyle\infty}{K} \Vert \leq O((\Delta_\phi^{-2} + m_l^{\frac{1}{2}} l) \kappa_\phi^2 m^{-\frac{1}{2}})) \geq 1-O(m^{-1})$ modulo logarithmic terms, where we denoted $\Delta_\phi = \frac{b^2}{a^2+b^2}$ and $\kappa_\phi = \frac{\vert a \vert + \vert b \vert}{\sqrt{a^2 + b^2}}$. This reveals in particular that the absolute value ($\Delta_\phi=1$, $\kappa_\phi=1$) beats the ReLU ($\Delta_\phi=\frac{1}{2}$, $\kappa_\phi=\sqrt{2}$) in terms of the concentration of the NTK.

Datasets play a critical role in medical imaging research, yet issues such as label quality, shortcuts, and metadata are often overlooked. This lack of attention may harm the generalizability of algorithms and, consequently, negatively impact patient outcomes. While existing medical imaging literature reviews mostly focus on machine learning (ML) methods, with only a few focusing on datasets for specific applications, these reviews remain static -- they are published once and not updated thereafter. This fails to account for emerging evidence, such as biases, shortcuts, and additional annotations that other researchers may contribute after the dataset is published. We refer to these newly discovered findings of datasets as research artifacts. To address this gap, we propose a living review that continuously tracks public datasets and their associated research artifacts across multiple medical imaging applications. Our approach includes a framework for the living review to monitor data documentation artifacts, and an SQL database to visualize the citation relationships between research artifact and dataset. Lastly, we discuss key considerations for creating medical imaging datasets, review best practices for data annotation, discuss the significance of shortcuts and demographic diversity, and emphasize the importance of managing datasets throughout their entire lifecycle. Our demo is publicly available at http://130.226.140.142.

The use of image analysis in automated photography management is an increasing trend in heritage institutions. Such tools alleviate the human cost associated with the manual and expensive annotation of new data sources while facilitating fast access to the citizenship through online indexes and search engines. However, available tagging and description tools are usually designed around modern photographs in English, neglecting historical corpora in minoritized languages, each of which exhibits intrinsic particularities. The primary objective of this research is to study the quantitative contribution of generative systems in the description of historical sources. This is done by contextualizing the task of captioning historical photographs from the Catalan archives as a case study. Our findings provide practitioners with tools and directions on transfer learning for captioning models based on visual adaptation and linguistic proximity.

Aim: To study the existence of subgroups by exploring the similarities between the attributes of the nodes of the groups, in relation to diet and gender and, to analyse the connectivity between groups based on aspects of similarities between them through SNA and artificial intelligence techniques. Methods: 235 students from 5 different educational centres participate in this study between March and December 2015. Data analysis carried out is divided into two blocks: social network analysis and unsupervised machine learning techniques. As for the social network analysis, the Girvan-Newman technique was applied to find the best number of cohesive groups within each of the friendship networks of the different classes analysed. Results: After applying Girvan-Newman in the three classes, the best division into clusters was respectively 2 for classroom A, 7 for classroom B and 6 for classroom C. There are significant differences between the groups and the gender and diet variables. After applying K-means using population diet as an input variable, a K-means clustering of 2 clusters for class A, 3 clusters for class B and 3 clusters for class C is obtained. Conclusion: Adolescents form subgroups within their classrooms. Subgroup cohesion is defined by the fact that nodes share similarities in aspects that influence obesity, they share attributes related to food quality and gender. The concept of homophily, related to SNA, justifies our results. Artificial intelligence techniques together with the application of the Girvan-Newman provide robustness to the structural analysis of similarities and cohesion between subgroups.

There's always been something seductive about a nanobot. Comic books and movies implore you to imagine these things, thousands of times thinner than a human hair and able to cruise around a body and repair a bone or heal an illness. Their scale is unfathomably finite. Their possibilities, sci-fi will have you believe, wildly infinite. While that incongruity makes it perfect for the denizens of a writers' room figuring out how to kill James Bond, it's also a sort of curse.

Spain's Prime Minister Pedro Sánchez closed the Spanish Strategy for R&D i in Artificial Intelligence workshop, held in Granada. During his speech, Sánchez highlighted that technologies related to artificial intelligence are already one of the main factors of growth, and hence Spain and Europe have to make a joint effort to move forward on this important line for social and economic progress. Pedro Sánchez explained that the document presented on Monday is the first step in drawing up the National Strategy on Artificial Intelligence, which 11 ministerial departments will work on and which will be ready later this year. Sánchez stressed the importance of science, innovation and universities for the present and future of the country. In this regard, he highlighted the creation of a specific ministerial department for these fields, the approval of a fundamental Royal Decree-Law to make the functioning of scientific bodies more flexible and the strengthening of equal opportunities, as well as the approval, last Friday, of the Research Personnel Statute on Training, the stabilisation of 1,500 temporary positions on public research bodies, which account for 10% of the total research workforce.