Institution profile

Universidad Autónoma de Madrid

Academic institutioneurope · es
Official website
Research library140linked papers
Opportunities0open roles
Selected work

Representative Papers

Statistical analysis of risk assessment factors and metrics to evaluate radicalisation in Twitter

Nov 01, 2017Future generations computer systems

This study addresses the problem of identifying radicalization risk among Weibo users. Methodologically, it proposes the first empirically grounded, multidimensional statistical framework for radicalization risk assessment tailored to real-world social media platforms. The framework integrates text feature engineering, social network behavior modeling, factor analysis, and correlation testing to systematically identify six interpretable risk indicators—including topic polarization degree, homophilous interaction rate, and keyword surge frequency. Its key contribution lies in establishing the first data-driven, multidimensional radicalization factor system and introducing a set of statistically significant, interpretable quantitative evaluation metrics. Evaluated on authentic Weibo data, the framework achieves 78.3% accuracy in detecting users’ radicalization tendencies. It thus provides a practical, deployable tool for platform-level content governance and early intervention.

41 citations2 influentialRead paper

Membership Inference Test: Auditing Training Data in Object Classification Models

Jan 19, 2026

This work proposes a Membership Inference Test (MINT) framework tailored for object detection models to address the risks of training data memorization and privacy leakage. By analyzing activation patterns in intermediate layers, the method integrates an object detector, an embedding extractor, and a customized MINT module to effectively determine whether a given input sample was part of the training set. Experiments on three public datasets—comprising over 174K images—demonstrate that the proposed approach achieves membership inference accuracy of 70%–80%. The study further identifies key factors influencing inference performance, such as the depth of the detection module’s input layer, thereby significantly enhancing the applicability and auditability of membership inference in complex vision tasks.

1 citationsRead paper

A Multimodal Dataset of Student Oral Presentations with Sensors and Evaluation Data

Jan 12, 2026

This work addresses the scarcity of ecologically valid multimodal datasets that hinders in-depth analysis of students’ oral presentation skills and the development of automated feedback systems. To bridge this gap, we introduce the SOPHIAS dataset, collected in authentic classroom settings, comprising 50 oral presentations and subsequent Q&A sessions delivered by 65 undergraduate students. The dataset integrates eight synchronized sensor modalities—including high-definition audio-video, eye-tracking, physiological signals (via smartwatches), interaction logs (keyboard, mouse, and clicker data), and presentation slides—and is annotated with standardized evaluations from instructors, peers, and self-assessments. Spanning approximately 12 hours of recordings, SOPHIAS is publicly available on GitHub and the Science Data Bank, offering the first high-ecological-validity benchmark resource for multimodal learning analytics, automated feedback generation, and peer assessment research.

1 citationsRead paper

MINT-Demo: Membership Inference Test Demonstrator

Mar 11, 2025

To address insufficient transparency in machine learning training data, this paper proposes and implements the first publicly demonstrable membership inference testing platform, enabling empirical determination of whether a specific sample was included in model training. Methodologically, we design a membership inference framework grounded in statistical significance testing and black-box model behavior analysis, systematically validated on a facial image dataset exceeding 22 million samples. By integrating heterogeneous facial data sources and mainstream recognition models, our framework supports cross-model generalization evaluation. Experiments achieve up to 89% membership identification accuracy across multiple publicly available face recognition models. This work represents the first engineering realization of membership inference as a reproducible, auditable open platform—establishing a novel paradigm for traceability and regulatory compliance verification in AI training processes.

1 citationsRead paper
Recent publications

Latest Papers