This study investigates the attribution patterns underlying Chain-of-Thought (CoT) reasoning in multilingual LLMs. While prior works demonstrate the role of CoT prompting in improving task performance, there are concerns regarding the faithfulness and interpretability of the generated reasoning chains. To assess these properties across languages, we applied two complementary attribution methods--ContextCite for step-level attribution and Inseq for token-level attribution--to the Qwen2.5 1.5B-Instruct model using the MGSM benchmark. Our experimental results highlight key findings such as: (1) attribution scores excessively emphasize the final reasoning step, particularly in incorrect generations; (2) structured CoT prompting significantly improves accuracy primarily for high-resource Latin-script languages; and (3) controlled perturbations via negation and distractor sentences reduce model accuracy and attribution coherence. These findings highlight the limitations of CoT prompting, particularly in terms of multilingual robustness and interpretive transparency.

Early identification of intensive care patients at risk of in-hospital mortality enables timely intervention and efficient resource allocation. Despite high predictive performance, existing machine learning approaches lack transparency and robustness, limiting clinical adoption. We present a lightweight, transparent multimodal ensemble that fuses physiological time-series measurements with unstructured clinical notes from the first 48 hours of an ICU stay. A logistic regression model combines predictions from two modality-specific models: a bidirectional LSTM for vitals and a finetuned ClinicalModernBERT transformer for notes. This traceable architecture allows for multilevel interpretability: feature attributions within each modality and direct per-case modality attributions quantifying how vitals and notes influence each decision. On the MIMIC-III benchmark, our late-fusion ensemble improves discrimination over the best single model (AUPRC 0.565 vs. 0.526; AUROC 0.891 vs. 0.876) while maintaining well-calibrated predictions. The system remains robust through a calibrated fallback when a modality is missing. These results demonstrate competitive performance with reliable, auditable risk estimates and transparent, predictable operation, which together are crucial for clinical use.

Quantifying numerical data involves addressing two key challenges: first, determining whether the data can be naturally quantified, and second, identifying the numerical intervals or ranges of values that correspond to specific value classes, referred to as "quantums," which represent statistically meaningful states. If such quantification is feasible, continuous streams of numerical data can be transformed into sequences of "symbols" that reflect the states of the system described by the measured parameter. People often perform this task intuitively, relying on common sense or practical experience, while information theory and computer science offer computable metrics for this purpose. In this study, we assess the applicability of metrics based on information compression and the Silhouette coefficient for quantifying numerical data. We also investigate the extent to which these metrics correlate with one another and with what is commonly referred to as "human intuition." Our findings suggest that the ability to classify numeric data values into distinct categories is associated with a Silhouette coefficient above 0.65 and a Dip Test below 0.5; otherwise, the data can be treated as following a unimodal normal distribution. Furthermore, when quantification is possible, the Silhouette coefficient appears to align more closely with human intuition than the "normalized centroid distance" method derived from information compression perspective.

The human brain undergoes continuous transformations across the lifespan, representing a natural component of aging that does not inherently signal pathological conditions [1]. Neurodegenerative disorders such as dementia can compromise the brain structure and accelerate aging processes. Understanding and characterizing healthy brain aging patterns therefore becomes essential for distinguishing normal aging from pathological neurodegeneration, potentially enabling earlier detection of neurodegenerative diseases. The Brain Age-Gap (BAG), i.e. the discrepancy between predicted brain age and chronological age, has emerged as a robust biomarker that captures pathological brain processes and offers insights into the rate at which an individual's brain ages in comparison to others in the population [2, 3]. It is not only associated with various neurological disorders, such as Alzheimer's disease, cognitive impairment, and Autism Spectrum Disorder, but also serves as an indicator of all-cause mortality [4, 5, 6, 7, 8] Brain age estimation has been approached through both conventional and machine learning techniques, analyzing either the whole brain, specific regions, or localized patches [9, 10, 11]. One particular study presented a method using T1-weighted MRI to predict age through region-level and voxel-level metrics [12]. Regression-based machine learning has shown promise for the brain age prediction, with kernel regression applied to whole-brain MRI across diverse age ranges [13]. Various algorithms including Support Vector Regression and Binary Decision Trees have been compared for their brain age prediction capabilities [14]. Additional regression techniques such as Relevance Vector Regression, Twin Support Vector Regression, and Gaussian Process Regression have been explored across different imaging modalities for age estimation and mortality prediction [11, 15, 16, 17].

This report introduces Kandinsky 5.0, a family of state-of-the-art foundation models for high-resolution image and 10-second video synthesis. The framework comprises three core line-up of models: Kandinsky 5.0 Image Lite - a line-up of 6B parameter image generation models, Kandinsky 5.0 Video Lite - a fast and lightweight 2B parameter text-to-video and image-to-video models, and Kandinsky 5.0 Video Pro - 19B parameter models that achieves superior video generation quality. We provide a comprehensive review of the data curation lifecycle - including collection, processing, filtering and clustering - for the multi-stage training pipeline that involves extensive pre-training and incorporates quality-enhancement techniques such as self-supervised fine-tuning (SFT) and reinforcement learning (RL)-based post-training. We also present novel architectural, training, and inference optimizations that enable Kandinsky 5.0 to achieve high generation speeds and state-of-the-art performance across various tasks, as demonstrated by human evaluation. As a large-scale, publicly available generative framework, Kandinsky 5.0 leverages the full potential of its pre-training and subsequent stages to be adapted for a wide range of generative applications. We hope that this report, together with the release of our open-source code and training checkpoints, will substantially advance the development and accessibility of high-quality generative models for the research community.

We study the $k$-means problem for a set $\mathcal{S} \subseteq \mathbb{R}^d$ of $n$ segments, aiming to find $k$ centers $X \subseteq \mathbb{R}^d$ that minimize $D(\mathcal{S},X) := \sum_{S \in \mathcal{S}} \min_{x \in X} D(S,x)$, where $D(S,x) := \int_{p \in S} |p - x| dp$ measures the total distance from each point along a segment to a center. Variants of this problem include handling outliers, employing alternative distance functions such as M-estimators, weighting distances to achieve balanced clustering, or enforcing unique cluster assignments. For any $\varepsilon > 0$, an $\varepsilon$-coreset is a weighted subset $C \subseteq \mathbb{R}^d$ that approximates $D(\mathcal{S},X)$ within a factor of $1 \pm \varepsilon$ for any set of $k$ centers, enabling efficient streaming, distributed, or parallel computation. We propose the first coreset construction that provably handles arbitrary input segments. For constant $k$ and $\varepsilon$, it produces a coreset of size $O(\log^2 n)$ computable in $O(nd)$ time. Experiments, including a real-time video tracking application, demonstrate substantial speedups with minimal loss in clustering accuracy, confirming both the practical efficiency and theoretical guarantees of our method.

In online convex optimization (OCO), a single loss function sequence is revealed over a time horizon of $T$, and an online algorithm has to choose its action at time $t$, before the loss function at time $t$ is revealed. The goal of the online algorithm is to incur minimal penalty (called $\textit{regret}$ compared to a static optimal action made by an optimal offline algorithm knowing all functions of the sequence in advance. In this paper, we broaden the horizon of OCO, and consider multi-objective OCO, where there are $K$ distinct loss function sequences, and an algorithm has to choose its action at time $t$, before the $K$ loss functions at time $t$ are revealed. To capture the tradeoff between tracking the $K$ different sequences, we consider the $\textit{min-max}$ regret, where the benchmark (optimal offline algorithm) takes a static action across all time slots that minimizes the maximum of the total loss (summed across time slots) incurred by each of the $K$ sequences. An online algorithm is allowed to change its action across time slots, and its {\it min-max} regret is defined as the difference between its $\textit{min-max}$ cost and that of the benchmark. The $\textit{min-max}$ regret is a stringent performance measure and an algorithm with small regret needs to `track' all loss function sequences closely at all times. We consider this $\textit{min-max}$ regret in the i.i.d. input setting where all loss functions are i.i.d. generated from an unknown distribution. For the i.i.d. model we propose a simple algorithm that combines the well-known $\textit{Hedge}$ and online gradient descent (OGD) and show via a remarkably simple proof that its expected $\textit{min-max}$ regret is $O(\sqrt{T \log K})$.

Traditional security scanners fail when facing new attack patterns they haven't seen before. They rely on fixed rules and predetermined signatures, making them blind to novel threats. We present a fundamentally different approach: instead of memorizing specific attack patterns, we learn what makes systems genuinely secure. Our key insight is simple yet powerful: context determines vulnerability. A SQL query that's safe in one environment becomes dangerous in another. By modeling this context-vulnerability relationship, we achieve something remarkable: our system detects attacks it has never seen before. We introduce context-aware verification that learns from genuine system behavior. Through reconstruction learning on secure systems, we capture their essential characteristics. When an unknown attack deviates from these patterns, our system recognizes it, even without prior knowledge of that specific attack type. We prove this capability theoretically, showing detection rates improve exponentially with context information I(W;C). Our framework combines three components: (1) reconstruction learning that models secure behavior, (2) multi-scale graph reasoning that aggregates contextual clues, and (3) attention mechanisms guided by reconstruction differences. Extensive experiments validate our approach: detection accuracy jumps from 58 percent to 82 percent with full context, unknown attack detection improves by 31 percent, and our system maintains above 90 percent accuracy even against completely novel attack vectors.

Satellite Earth-observation (EO) time series in the optical and microwave ranges of the electromagnetic spectrum are often irregular due to orbital patterns and cloud obstruction. Compositing addresses these issues but loses information with respect to vegetation phenology, which is critical for many downstream tasks. Instead, we present TESSERA, a pixel-wise foundation model for multi-modal (Sentinel-1/2) EO time series that learns robust, label-efficient em-beddings. During model training, TESSERA uses Barlow Twins and sparse random temporal sampling to enforce invariance to the selection of valid observations. W e employ two key regularizers: global shuffling to decorrelate spatial neighborhoods and mix-based regulation to improve invariance under extreme sparsity. W e find that for diverse classification, segmentation, and regression tasks, TESSERA embeddings deliver state-of-the-art accuracy with high label efficiency, often requiring only a small task head and minimal computation. T o democratize access, adhere to F AIR principles, and simplify use, we release global, annual, 10m, pixel-wise int8 embeddings together with open weights/code and lightweight adaptation heads, thus providing practical tooling for large-scale retrieval and inference at planetary scale. The model training/inference code, downstream task code, and pre-generated embeddings can be accessed at https://github.com/ucam-eo.

Estimation of causal effects is critical to a range of scientific disciplines. Existing methods for this task either require interventional data, knowledge about the ground truth causal graph, or rely on assumptions such as unconfoundedness, restricting their applicability in real-world settings. In the domain of tabular machine learning, Prior-data fitted networks (PFNs) have achieved state-of-the-art predictive performance, having been pre-trained on synthetic data to solve tabular prediction problems via in-context learning. To assess whether this can be transferred to the harder problem of causal effect estimation, we pre-train PFNs on synthetic data drawn from a wide variety of causal structures, including interventions, to predict interventional outcomes given observational data. Through extensive experiments on synthetic case studies, we show that our approach allows for the accurate estimation of causal effects without knowledge of the underlying causal graph. We also perform ablation studies that elucidate Do-PFN's scalability and robustness across datasets with a variety of causal characteristics.