Hybrid Pooling with LLMs via Relevance Context Learning | Proceedings of the 49th International ACM SIGIR Conference on Research and Development in Information Retrieval

doi.org

Amparo Alonso, catedrática en la Universidade da Coruña y Premio Nacional de Informática: 'La regulación es imprescindible'

laopinioncoruna.es

Reliable hierarchical operating system fingerprinting via conformal prediction - International Journal of Information Security

Operating System (OS) fingerprinting is critical for network security, but conventional methods do not provide formal uncertainty quantification mechanisms. Conformal Prediction (CP) could be directly wrapped around existing methods to obtain prediction sets with guaranteed coverage. However, a direct application of CP would treat OS identification as a flat classification problem, ignoring the natural taxonomic structure of OSs and providing brittle point predictions. This work addresses these limitations by introducing and evaluating two distinct structured CP strategies: level-wise CP (L-CP), which calibrates each hierarchy level independently, and projection-based CP (P-CP), which ensures structural consistency by projecting leaf-level sets upwards. Our results demonstrate that, while both methods satisfy validity guarantees, they expose a fundamental trade-off between level-wise efficiency and structural consistency. L-CP yields tighter prediction sets suitable for human forensic analysis but suffers from taxonomic inconsistencies. Conversely, P-CP guarantees hierarchically consistent, nested sets ideal for automated policy enforcement, albeit at the cost of reduced efficiency at coarser levels.

doi.org

La ingeniera que programa por la igualdad

farodevigo.es

On how the rules that extend the concede-and-divide principle differ for pairs of claimants - TOP

In the context of conflicting claims problems, the Talmud, adjusted proportional, random arrival, minimal overlap, and average-of-awards rules are well-known extensions of the two-claimant concede-and-divide rule to larger populations. We analyze how the awards recommended by these rules for each couple of claimants differ from what they would receive if the concede-and-divide principle were applied to distribute among them their joint payoff. To compare the behavior of these rules for pairs of claimants, we rely on geometric features of the corresponding two-dimensional paths of awards. We study properties pertaining to the impact, in absolute and relative terms, that one claimant leaving the problem has on the awards of the remaining claimants.

doi.org

3D OCT-Based Retinal Biomarker Analysis for Automatic Regional-Wise Characterization of Neurodegenerative Diseases - Journal of Imaging Informatics in Medicine

Neurodegenerative diseases (NDDs) such as Alzheimer’s disease (AD), essential tremor (ET), multiple sclerosis (MS), and Parkinson’s disease (PD) are complex disorders that often exhibit overlapping symptoms, leading to diagnostic challenges. Given the increasing interest in retinal imaging as a non-invasive biomarker for neurodegeneration, this study proposes a fully automated machine learning pipeline for disease characterization using optical coherence tomography (OCT). We analyze macular thickness patterns across three key and relevant retinal elements: retinal nerve fibre layer (RNFL), ganglion cell layer to Bruch’s membrane (GCL-BM), and the total retina. These are processed by two complementary regional layouts: the standard ETDRS scheme and a custom 3 $$\times $$ × 3 quadrant grid. These measurements are used to train multiple classifiers to distinguish between healthy controls and NDDs either collectively or individually. The proposed method processes 34,375 OCT B-scans from 353 subjects and highlights disease-specific thickness patterns with a pathological distinction score ranging up to 0.71 depending on the retinal region, disease, and classifier. Sector-based grids generally outperform quadrant-based ones, revealing highly localized pathological signatures. Our findings demonstrate that each disease manifests distinct retinal alterations, aligning with current clinical literature while offering novel insights for ET and PD. The study reinforces the potential of grid-based OCT analysis as a discriminative and fully automatic screening tool, paving the way for improved early diagnosis and differential analysis of NDDs through retinal biomarkers.

doi.org

Multilevel representations of multiple geographical objects

Abstract. In this work, we explore different solutions for the problem of representing and querying geographical objects at different scales (zoom levels).

doi.org

A study of word embedding models for measuring topic coherence - Knowledge and Information Systems

Topic modeling has emerged as a crucial tool in the field of natural language processing, enabling the automatic discovery of latent structures in large textual corpora. However, determining the quality of the topics remains a significant challenge, particularly in measuring the coherence of the top words of the extracted topics. Early efforts relied on human judgments, but these approaches are resource-intensive. Automated coherence metrics have since been developed. For example, some measures exploit word co-occurrence, while other methods are grounded in distributional semantics (e.g., employing word embeddings). In this study, we thoroughly explore the application of embedded representations to evaluate the quality of topics. While a number of isolated studies have analyzed the role of specific word representation techniques for measuring topic coherence, a complete picture of their effectiveness is still lacking. This work brings together different embedding-based approaches, including Word2Vec, FastText, GloVe, and BERT, which had been studied separately, and extends prior research by incorporating additional models, such as RoBERTa, ALBERT and MPNET. Topic coherence is measured by computing similarity scores between word embeddings, thus obtaining rich semantic associations that traditional measures may overlook. Our analysis demonstrates that these methods are as effective as, and often surpass, classical coherence measures. Our results contribute to a growing body of research advocating for advanced semantic representations as robust alternatives to traditional approaches in evaluating topic model coherence.

doi.org