Aythami Morales

Miguel Lopez-Duran, Elena Marrero, Julian Fierrez et al.

Document Visual Question Answering (DocVQA) presents a complex multimodal challenge, requiring models to exploit visual, textual, and layout information from documents. Although Vision-Language Models (VLMs) have shown remarkable performance in text-vision tasks, their robustness and transferability to different document domains remains underexplored. In this study, we present a comprehensive evaluation of 8 open-source pretrained VLMs on DocVQA in three different document domains: industrial documents of varying type, infographics, and presentation slides. We systematically assess model performance under zero-shot evaluations, fully supervised finetuning with inter- and intra-dataset evaluations, and few-shot learning evaluations of knowledge transfer between domains. Our findings demonstrate that while large pretrained VLMs possess strong zero-shot baselines for structured layouts, their performance strongly decreases on visually complex layouts of infographics and slides. Although parameter scaling is a dominant factor on performance, supervised finetuning yields higher relative gains in smaller architectures. Furthermore, our cross-domain and few-shot experiments show that visual understanding is the main bottleneck for DocVQA, not a lack of knowledge from the VLMs. Using 50 target domain samples, the models finetuned in DocVQA with datasets of different domains rapidly adapt to the target domain documents, even surpassing their fully supervised counterparts in some cases.

Miguel Lopez-Duran, Julian Fierrez, Aythami Morales et al.

We present CrimeNER Demo, an AI-powered platform that enables us to extract general crime-related information from documents and classify them into entity types with two levels of granularity. We provide pretrained NER models on the CrimeNER database, and we give the possibility to users to provide their own annotated data to train models for their own specific cases. This demonstrator aims to promote crime-related NER research and provides a practical tool to automatically extract crime information for researchers and law enforcement agencies. The demonstrator includes: i) Pretrained NER models on the crime domain; ii) Possibility to finetune the models on specific data annotated by the user; and iii) An automatic pipeline to extract and annotate crime entities from documents. The demo platform, a tutorial to run the demo, and a video demonstration are publicly available on GitHub.

Ignacio Serna, Aythami Morales, Julian Fierrez

This study investigates the presence and propagation of bias within Neural Networks through a comprehensive multi-level analysis spanning the learned latent space, layer activations, and the network's parameters. Based on this taxonomy, we propose three bias detection approaches: 1) SpaceBias (new method), which characterizes the latent space prior to the final classification layer using neighbor-probability distributions and quantifies bias with the two-sample Kolmogorov-Smirnov test on the per-group distributions. 2) ActivationBias (extension of the existing method InsideBias), which analyzes the activations of neural network filters and quantifies bias via a Mann-Whitney U test, based on the observed fact that underrepresented groups exhibit lower activation levels in the final convolutional layers. 3) WeightBias (extension of the existing method IFBiD), which uses a secondary neural network trained to identify biased patterns directly in the parameters of task-specific models. Unlike conventional methods, which assess neural network outcomes and treat the model as a black box, our proposed techniques provide insight into how biases manifest within the network architecture itself at different levels, offering a more nuanced and detailed understanding. Experiments are conducted on two complementary applications: gender classification in the DiveFace dataset (72,000 face images) and digit classification on a colored-MNIST benchmark with controlled bias severity. In total, more than 127,000 models with varying degrees and types of bias were trained and evaluated. The severity sweep shows that the internal disparity, and with it the detection performance, decreases smoothly as the training distribution approaches balance. The results highlight the importance of methods that provide deeper insight into the behavior of AI models.

Alejandro Peña, Ignacio Serna, Aythami Morales et al.

The presence of decision-making algorithms in society is rapidly increasing nowadays, while concerns about their transparency and the possibility of these algorithms becoming new sources of discrimination are arising. There is a certain consensus about the need to develop AI applications with a Human-Centric approach. Human-Centric Machine Learning needs to be developed based on four main requirements: (i) utility and social good; (ii) privacy and data ownership; (iii) transparency and accountability; and (iv) fairness in AI-driven decision-making processes. All these four Human-Centric requirements are closely related to each other. With the aim of studying how current multimodal algorithms based on heterogeneous sources of information are affected by sensitive elements and inner biases in the data, we propose a fictitious case study focused on automated recruitment: FairCVtest. We train automatic recruitment algorithms using a set of multimodal synthetic profiles including image, text, and structured data, which are consciously scored with gender and racial biases. FairCVtest shows the capacity of the Artificial Intelligence (AI) behind automatic recruitment tools built this way (a common practice in many other application scenarios beyond recruitment) to extract sensitive information from unstructured data and exploit it in combination to data biases in undesirable (unfair) ways. We present an overview of recent works developing techniques capable of removing sensitive information and biases from the decision-making process of deep learning architectures, as well as commonly used databases for fairness research in AI. We demonstrate how learning approaches developed to guarantee privacy in latent spaces can lead to unbiased and fair automatic decision-making process.

Alejandro Peña, Aythami Morales, Julian Fierrez et al.

The analysis of public affairs documents is crucial for citizens as it promotes transparency, accountability, and informed decision-making. It allows citizens to understand government policies, participate in public discourse, and hold representatives accountable. This is crucial, and sometimes a matter of life or death, for companies whose operation depend on certain regulations. Large Language Models (LLMs) have the potential to greatly enhance the analysis of public affairs documents by effectively processing and understanding the complex language used in such documents. In this work, we analyze the performance of LLMs in classifying public affairs documents. As a natural multi-label task, the classification of these documents presents important challenges. In this work, we use a regex-powered tool to collect a database of public affairs documents with more than 33K samples and 22.5M tokens. Our experiments assess the performance of 4 different Spanish LLMs to classify up to 30 different topics in the data in different configurations. The results shows that LLMs can be of great use to process domain-specific documents, such as those in the domain of public affairs.

Roberto Daza, Aythami Morales, Ruben Tolosana et al.

We present edBB-Demo, a demonstrator of an AI-powered research platform for student monitoring in remote education. The edBB platform aims to study the challenges associated to user recognition and behavior understanding in digital platforms. This platform has been developed for data collection, acquiring signals from a variety of sensors including keyboard, mouse, webcam, microphone, smartwatch, and an Electroencephalography band. The information captured from the sensors during the student sessions is modelled in a multimodal learning framework. The demonstrator includes: i) Biometric user authentication in an unsupervised environment; ii) Human action recognition based on remote video analysis; iii) Heart rate estimation from webcam video; and iv) Attention level estimation from facial expression analysis.

Pietro Melzi, Ruben Tolosana, Ruben Vera-Rodriguez et al.

Despite the widespread adoption of face recognition technology around the world, and its remarkable performance on current benchmarks, there are still several challenges that must be covered in more detail. This paper offers an overview of the Face Recognition Challenge in the Era of Synthetic Data (FRCSyn) organized at WACV 2024. This is the first international challenge aiming to explore the use of synthetic data in face recognition to address existing limitations in the technology. Specifically, the FRCSyn Challenge targets concerns related to data privacy issues, demographic biases, generalization to unseen scenarios, and performance limitations in challenging scenarios, including significant age disparities between enrollment and testing, pose variations, and occlusions. The results achieved in the FRCSyn Challenge, together with the proposed benchmark, contribute significantly to the application of synthetic data to improve face recognition technology.

Giuseppe Stragapede, Ruben Vera-Rodriguez, Ruben Tolosana et al.

Current mobile user authentication systems based on PIN codes, fingerprint, and face recognition have several shortcomings. Such limitations have been addressed in the literature by exploring the feasibility of passive authentication on mobile devices through behavioral biometrics. In this line of research, this work carries out a comparative analysis of unimodal and multimodal behavioral biometric traits acquired while the subjects perform different activities on the phone such as typing, scrolling, drawing a number, and tapping on the screen, considering the touchscreen and the simultaneous background sensor data (accelerometer, gravity sensor, gyroscope, linear accelerometer, and magnetometer). Our experiments are performed over HuMIdb, one of the largest and most comprehensive freely available mobile user interaction databases to date. A separate Recurrent Neural Network (RNN) with triplet loss is implemented for each single modality. Then, the weighted fusion of the different modalities is carried out at score level. In our experiments, the most discriminative background sensor is the magnetometer, whereas among touch tasks the best results are achieved with keystroke in a fixed-text scenario. In all cases, the fusion of modalities is very beneficial, leading to Equal Error Rates (EER) ranging from 4% to 9% depending on the modality combination in a 3-second interval.

Daniel DeAlcala, Ignacio Serna, Aythami Morales et al.

The new regulatory framework proposal on Artificial Intelligence (AI) published by the European Commission establishes a new risk-based legal approach. The proposal highlights the need to develop adequate risk assessments for the different uses of AI. This risk assessment should address, among others, the detection and mitigation of bias in AI. In this work we analyze statistical approaches to measure biases in automatic decision-making systems. We focus our experiments in face recognition technologies. We propose a novel way to measure the biases in machine learning models using a statistical approach based on the N-Sigma method. N-Sigma is a popular statistical approach used to validate hypotheses in general science such as physics and social areas and its application to machine learning is yet unexplored. In this work we study how to apply this methodology to develop new risk assessment frameworks based on bias analysis and we discuss the main advantages and drawbacks with respect to other popular statistical tests.

Giuseppe Stragapede, Ruben Vera-Rodriguez, Ruben Tolosana et al.

Mobile behavioral biometrics have become a popular topic of research, reaching promising results in terms of authentication, exploiting a multimodal combination of touchscreen and background sensor data. However, there is no way of knowing whether state-of-the-art classifiers in the literature can distinguish between the notion of user and device. In this article, we present a new database, BehavePassDB, structured into separate acquisition sessions and tasks to mimic the most common aspects of mobile Human-Computer Interaction (HCI). BehavePassDB is acquired through a dedicated mobile app installed on the subjects' devices, also including the case of different users on the same device for evaluation. We propose a standard experimental protocol and benchmark for the research community to perform a fair comparison of novel approaches with the state of the art. We propose and evaluate a system based on Long-Short Term Memory (LSTM) architecture with triplet loss and modality fusion at score level.

Sergio Romero-Tapiador, Ruben Tolosana, Aythami Morales et al.

Nowadays millions of images are shared on social media and web platforms. In particular, many of them are food images taken from a smartphone over time, providing information related to the individual's diet. On the other hand, eating behaviours are directly related to some of the most prevalent diseases in the world. Exploiting recent advances in image processing and Artificial Intelligence (AI), this scenario represents an excellent opportunity to: i) create new methods that analyse the individuals' health from what they eat, and ii) develop personalised recommendations to improve nutrition and diet under specific circumstances (e.g., obesity or COVID). Having tunable tools for creating food image datasets that facilitate research in both lines is very much needed. This paper proposes AI4Food-NutritionFW, a framework for the creation of food image datasets according to configurable eating behaviours. AI4Food-NutritionFW simulates a user-friendly and widespread scenario where images are taken using a smartphone. In addition to the framework, we also provide and describe a unique food image dataset that includes 4,800 different weekly eating behaviours from 15 different profiles and 1,200 subjects. Specifically, we consider profiles that comply with actual lifestyles from healthy eating behaviours (according to established knowledge), variable profiles (e.g., eating out, holidays), to unhealthy ones (e.g., excess of fast food or sweets). Finally, we automatically evaluate a healthy index of the subject's eating behaviours using multidimensional metrics based on guidelines for healthy diets proposed by international organisations, achieving promising results (99.53% and 99.60% accuracy and sensitivity, respectively). We also release to the research community a software implementation of our proposed AI4Food-NutritionFW and the mentioned food image dataset created with it.

Roberto Daza, Luis F. Gomez, Aythami Morales et al.

This work presents a new multimodal system for remote attention level estimation based on multimodal face analysis. Our multimodal approach uses different parameters and signals obtained from the behavior and physiological processes that have been related to modeling cognitive load such as faces gestures (e.g., blink rate, facial actions units) and user actions (e.g., head pose, distance to the camera). The multimodal system uses the following modules based on Convolutional Neural Networks (CNNs): Eye blink detection, head pose estimation, facial landmark detection, and facial expression features. First, we individually evaluate the proposed modules in the task of estimating the student's attention level captured during online e-learning sessions. For that we trained binary classifiers (high or low attention) based on Support Vector Machines (SVM) for each module. Secondly, we find out to what extent multimodal score level fusion improves the attention level estimation. The mEBAL database is used in the experimental framework, a public multi-modal database for attention level estimation obtained in an e-learning environment that contains data from 38 users while conducting several e-learning tasks of variable difficulty (creating changes in student cognitive loads).

Roberto Daza, Aythami Morales, Julian Fierrez et al.

This work introduces a new multispectral database and novel approaches for eyeblink detection in RGB and Near-Infrared (NIR) individual images. Our contributed dataset (mEBAL2, multimodal Eye Blink and Attention Level estimation, Version 2) is the largest existing eyeblink database, representing a great opportunity to improve data-driven multispectral approaches for blink detection and related applications (e.g., attention level estimation and presentation attack detection in face biometrics). mEBAL2 includes 21,100 image sequences from 180 different students (more than 2 million labeled images in total) while conducting a number of e-learning tasks of varying difficulty or taking a real course on HTML initiation through the edX MOOC platform. mEBAL2 uses multiple sensors, including two Near-Infrared (NIR) and one RGB camera to capture facial gestures during the execution of the tasks, as well as an Electroencephalogram (EEG) band to get the cognitive activity of the user and blinking events. Furthermore, this work proposes a Convolutional Neural Network architecture as benchmark for blink detection on mEBAL2 with performances up to 97%. Different training methodologies are implemented using the RGB spectrum, NIR spectrum, and the combination of both to enhance the performance on existing eyeblink detectors. We demonstrate that combining NIR and RGB images during training improves the performance of RGB eyeblink detectors (i.e., detection based only on a RGB image). Finally, the generalization capacity of the proposed eyeblink detectors is validated in wilder and more challenging environments like the HUST-LEBW dataset to show the usefulness of mEBAL2 to train a new generation of data-driven approaches for eyeblink detection.

Sergio Romero-Tapiador, Ruben Tolosana, Aythami Morales et al.

Maintaining a healthy lifestyle has become increasingly challenging in today's sedentary society marked by poor eating habits. To address this issue, both national and international organisations have made numerous efforts to promote healthier diets and increased physical activity. However, implementing these recommendations in daily life can be difficult, as they are often generic and not tailored to individuals. This study presents the AI4Food-NutritionDB database, the first nutrition database that incorporates food images and a nutrition taxonomy based on recommendations by national and international health authorities. The database offers a multi-level categorisation, comprising 6 nutritional levels, 19 main categories (e.g., "Meat"), 73 subcategories (e.g., "White Meat"), and 893 specific food products (e.g., "Chicken"). The AI4Food-NutritionDB opens the doors to new food computing approaches in terms of food intake frequency, quality, and categorisation. Also, we present a standardised experimental protocol and benchmark including three tasks based on the nutrition taxonomy (i.e., category, subcategory, and final product recognition). These resources are available to the research community, including our deep learning models trained on AI4Food-NutritionDB, which can serve as pre-trained models, achieving accurate recognition results for challenging food image databases.

Mahdi Ghafourian, Julian Fierrez, Luis Felipe Gomez et al.

The remarkable success of face recognition (FR) has endangered the privacy of internet users particularly in social media. Recently, researchers turned to use adversarial examples as a countermeasure. In this paper, we assess the effectiveness of using two widely known adversarial methods (BIM and ILLC) for de-identifying personal images. We discovered, unlike previous claims in the literature, that it is not easy to get a high protection success rate (suppressing identification rate) with imperceptible adversarial perturbation to the human visual system. Finally, we found out that the transferability of adversarial examples is highly affected by the training parameters of the network with which they are generated.

Giuseppe Stragapede, Ruben Vera-Rodriguez, Ruben Tolosana et al.

This paper describes the experimental framework and results of the IJCB 2022 Mobile Behavioral Biometrics Competition (MobileB2C). The aim of MobileB2C is benchmarking mobile user authentication systems based on behavioral biometric traits transparently acquired by mobile devices during ordinary Human-Computer Interaction (HCI), using a novel public database, BehavePassDB, and a standard experimental protocol. The competition is divided into four tasks corresponding to typical user activities: keystroke, text reading, gallery swiping, and tapping. The data are composed of touchscreen data and several background sensor data simultaneously acquired. "Random" (different users with different devices) and "skilled" (different user on the same device attempting to imitate the legitimate one) impostor scenarios are considered. The results achieved by the participants show the feasibility of user authentication through behavioral biometrics, although this proves to be a non-trivial challenge. MobileB2C will be established as an on-going competition.

Roberto Daza, Luis F. Gomez, Julian Fierrez et al.

This work introduces an innovative method for estimating attention levels (cognitive load) using an ensemble of facial analysis techniques applied to webcam videos. Our method is particularly useful, among others, in e-learning applications, so we trained, evaluated, and compared our approach on the mEBAL2 database, a public multi-modal database acquired in an e-learning environment. mEBAL2 comprises data from 60 users who performed 8 different tasks. These tasks varied in difficulty, leading to changes in their cognitive loads. Our approach adapts state-of-the-art facial analysis technologies to quantify the users' cognitive load in the form of high or low attention. Several behavioral signals and physiological processes related to the cognitive load are used, such as eyeblink, heart rate, facial action units, and head pose, among others. Furthermore, we conduct a study to understand which individual features obtain better results, the most efficient combinations, explore local and global features, and how temporary time intervals affect attention level estimation, among other aspects. We find that global facial features are more appropriate for multimodal systems using score-level fusion, particularly as the temporal window increases. On the other hand, local features are more suitable for fusion through neural network training with score-level fusion approaches. Our method outperforms existing state-of-the-art accuracies using the public mEBAL2 benchmark.

Giuseppe Stragapede, Ruben Vera-Rodriguez, Ruben Tolosana et al.

Analyzing keystroke dynamics (KD) for biometric verification has several advantages: it is among the most discriminative behavioral traits; keyboards are among the most common human-computer interfaces, being the primary means for users to enter textual data; its acquisition does not require additional hardware, and its processing is relatively lightweight; and it allows for transparently recognizing subjects. However, the heterogeneity of experimental protocols and metrics, and the limited size of the databases adopted in the literature impede direct comparisons between different systems, thus representing an obstacle in the advancement of keystroke biometrics. To alleviate this aspect, we present a new experimental framework to benchmark KD-based biometric verification performance and fairness based on tweet-long sequences of variable transcript text from over 185,000 subjects, acquired through desktop and mobile keyboards, extracted from the Aalto Keystroke Databases. The framework runs on CodaLab in the form of the Keystroke Verification Challenge (KVC). Moreover, we also introduce a novel fairness metric, the Skewed Impostor Ratio (SIR), to capture inter- and intra-demographic group bias patterns in the verification scores. We demonstrate the usefulness of the proposed framework by employing two state-of-the-art keystroke verification systems, TypeNet and TypeFormer, to compare different sets of input features, achieving a less privacy-invasive system, by discarding the analysis of text content (ASCII codes of the keys pressed) in favor of extended features in the time domain. Our experiments show that this approach allows to maintain satisfactory performance.

Giuseppe Stragapede, Paula Delgado-Santos, Ruben Tolosana et al.

The broad usage of mobile devices nowadays, the sensitiveness of the information contained in them, and the shortcomings of current mobile user authentication methods are calling for novel, secure, and unobtrusive solutions to verify the users' identity. In this article, we propose TypeFormer, a novel Transformer architecture to model free-text keystroke dynamics performed on mobile devices for the purpose of user authentication. The proposed model consists in Temporal and Channel Modules enclosing two Long Short-Term Memory (LSTM) recurrent layers, Gaussian Range Encoding (GRE), a multi-head Self-Attention mechanism, and a Block-Recurrent structure. Experimenting on one of the largest public databases to date, the Aalto mobile keystroke database, TypeFormer outperforms current state-of-the-art systems achieving Equal Error Rate (EER) values of 3.25% using only 5 enrolment sessions of 50 keystrokes each. In such way, we contribute to reducing the traditional performance gap of the challenging mobile free-text scenario with respect to its desktop and fixed-text counterparts. Additionally, we analyse the behaviour of the model with different experimental configurations such as the length of the keystroke sequences and the amount of enrolment sessions, showing margin for improvement with more enrolment data. Finally, a cross-database evaluation is carried out, demonstrating the robustness of the features extracted by TypeFormer in comparison with existing approaches.

Mahdi Ghafourian, Julian Fierrez, Ruben Vera-Rodriguez et al.

Cancelable biometrics are a group of techniques to transform the input biometric to an irreversible feature intentionally using a transformation function and usually a key in order to provide security and privacy in biometric recognition systems. This transformation is repeatable enabling subsequent biometric comparisons. This paper is introducing a new idea to exploit as a transformation function for cancelable biometrics aimed at protecting the templates against iterative optimization attacks. Our proposed scheme is based on time-varying keys (random biometrics in our case) and morphing transformations. An experimental implementation of the proposed scheme is given for face biometrics. The results confirm that the proposed approach is able to withstand against leakage attacks while improving the recognition performance.

Giuseppe Stragapede, Paula Delgado-Santos, Ruben Tolosana et al.

Among user authentication methods, behavioural biometrics has proven to be effective against identity theft as well as user-friendly and unobtrusive. One of the most popular traits in the literature is keystroke dynamics due to the large deployment of computers and mobile devices in our society. This paper focuses on improving keystroke biometric systems on the free-text scenario. This scenario is characterised as very challenging due to the uncontrolled text conditions, the influence of the user's emotional and physical state, and the in-use application. To overcome these drawbacks, methods based on deep learning such as Convolutional Neural Networks (CNNs) and Recurrent Neural Networks (RNNs) have been proposed in the literature, outperforming traditional machine learning methods. However, these architectures still have aspects that need to be reviewed and improved. To the best of our knowledge, this is the first study that proposes keystroke biometric systems based on Transformers. The proposed Transformer architecture has achieved Equal Error Rate (EER) values of 3.84\% in the popular Aalto mobile keystroke database using only 5 enrolment sessions, outperforming by a large margin other state-of-the-art approaches in the literature.

Luis F. Gomez, Julian Fierrez, Aythami Morales et al.

Presentation Attack Detection (PAD) is a crucial stage in facial recognition systems to avoid leakage of personal information or spoofing of identity to entities. Recently, pulse detection based on remote photoplethysmography (rPPG) has been shown to be effective in face presentation attack detection. This work presents three different approaches to the presentation attack detection based on rPPG: (i) The physiological domain, a domain using rPPG-based models, (ii) the Deepfakes domain, a domain where models were retrained from the physiological domain to specific Deepfakes detection tasks; and (iii) a new Presentation Attack domain was trained by applying transfer learning from the two previous domains to improve the capability to differentiate between bona-fides and attacks. The results show the efficiency of the rPPG-based models for presentation attack detection, evidencing a 21.70% decrease in average classification error rate (ACER) (from 41.03% to 19.32%) when the presentation attack domain is compared to the physiological and Deepfakes domains. Our experiments highlight the efficiency of transfer learning in rPPG-based models and perform well in presentation attack detection in instruments that do not allow copying of this physiological feature.

Daniel DeAlcala, Aythami Morales, Ruben Tolosana et al.

This work proposes a data driven learning model for the synthesis of keystroke biometric data. The proposed method is compared with two statistical approaches based on Universal and User-dependent models. These approaches are validated on the bot detection task, using the keystroke synthetic data to improve the training process of keystroke-based bot detection systems. Our experimental framework considers a dataset with 136 million keystroke events from 168 thousand subjects. We have analyzed the performance of the three synthesis approaches through qualitative and quantitative experiments. Different bot detectors are considered based on several supervised classifiers (Support Vector Machine, Random Forest, Gaussian Naive Bayes and a Long Short-Term Memory network) and a learning framework including human and synthetic samples. The experiments demonstrate the realism of the synthetic samples. The classification results suggest that in scenarios with large labeled data, these synthetic samples can be detected with high accuracy. However, in few-shot learning scenarios it represents an important challenge. Furthermore, these results show the great potential of the presented models.

Sergio Romero-Tapiador, Ruben Tolosana, Aythami Morales et al.

Early detection of chronic and Non-Communicable Diseases (NCDs) is crucial for effective treatment during the initial stages. This study explores the application of wearable devices and Artificial Intelligence (AI) in order to predict weight loss changes in overweight and obese individuals. Using wearable data from a 1-month trial involving around 100 subjects from the AI4FoodDB database, including biomarkers, vital signs, and behavioral data, we identify key differences between those achieving weight loss (>= 2% of their initial weight) and those who do not. Feature selection techniques and classification algorithms reveal promising results, with the Gradient Boosting classifier achieving 84.44% Area Under the Curve (AUC). The integration of multiple data sources (e.g., vital signs, physical and sleep activity, etc.) enhances performance, suggesting the potential of wearable devices and AI in personalized healthcare.

Alejandro Peña, Aythami Morales, Julian Fierrez et al.

Every day, thousands of digital documents are generated with useful information for companies, public organizations, and citizens. Given the impossibility of processing them manually, the automatic processing of these documents is becoming increasingly necessary in certain sectors. However, this task remains challenging, since in most cases a text-only based parsing is not enough to fully understand the information presented through different components of varying significance. In this regard, Document Layout Analysis (DLA) has been an interesting research field for many years, which aims to detect and classify the basic components of a document. In this work, we used a procedure to semi-automatically annotate digital documents with different layout labels, including 4 basic layout blocks and 4 text categories. We apply this procedure to collect a novel database for DLA in the public affairs domain, using a set of 24 data sources from the Spanish Administration. The database comprises 37.9K documents with more than 441K document pages, and more than 8M labels associated to 8 layout block units. The results of our experiments validate the proposed text labeling procedure with accuracy up to 99%.

Mahdi Ghafourian, Julian Fierrez, Ruben Vera-Rodriguez et al.

Federated Learning (FL) is a machine learning paradigm to conduct collaborative learning among clients on a joint model. The primary goal is to share clients' local training parameters with an integrating server while preserving their privacy. This method permits to exploit the potential of massive mobile users' data for the benefit of machine learning models' performance while keeping sensitive data on local devices. On the downside, FL raises security and privacy concerns that have just started to be studied. To address some of the key threats in FL, researchers have proposed to use secure aggregation methods (e.g. homomorphic encryption, secure multiparty computation, etc.). These solutions improve some security and privacy metrics, but at the same time bring about other serious threats such as poisoning attacks, backdoor attacks, and free running attacks. This paper proposes a new defense method against poisoning attacks in FL called SaFL (Sybil-aware Federated Learning) that minimizes the effect of sybils with a novel time-variant aggregation scheme.

Imanol Solano, Alejandro Peña, Aythami Morales et al.

We present a novel metric designed, among other applications, to quantify biased behaviors of machine learning models. As its core, the metric consists of a new similarity metric between score distributions that balances both their general shapes and tails' probabilities. In that sense, our proposed metric may be useful in many application areas. Here we focus on and apply it to the operational evaluation of face recognition systems, with special attention to quantifying demographic biases; an application where our metric is especially useful. The topic of demographic bias and fairness in biometric recognition systems has gained major attention in recent years. The usage of these systems has spread in society, raising concerns about the extent to which these systems treat different population groups. A relevant step to prevent and mitigate demographic biases is first to detect and quantify them. Traditionally, two approaches have been studied to quantify differences between population groups in machine learning literature: 1) measuring differences in error rates, and 2) measuring differences in recognition score distributions. Our proposed Comprehensive Equity Index (CEI) trade-offs both approaches combining both errors from distribution tails and general distribution shapes. This new metric is well suited to real-world scenarios, as measured on NIST FRVT evaluations, involving high-performance systems and realistic face databases including a wide range of covariates and demographic groups. We first show the limitations of existing metrics to correctly assess the presence of biases in realistic setups and then propose our new metric to tackle these limitations. We tested the proposed metric with two state-of-the-art models and four widely used databases, showing its capacity to overcome the main flaws of previous bias metrics.

Gonzalo Garrido-Lopez, Luis F. Gomez, Julian Fierrez et al.

Sprinting is a determinant ability, especially in team sports. The kinematics of the sprint have been studied in the past using different methods specially developed considering human biomechanics and, among those methods, markerless systems stand out as very cost-effective. On the other hand, we have now multiple general methods for pixel and body tracking based on recent machine learning breakthroughs with excellent performance in body tracking, but these excellent trackers do not generally consider realistic human biomechanics. This investigation first adapts two of these general trackers (MoveNet and CoTracker) for realistic biomechanical analysis and then evaluate them in comparison to manual tracking (with key points manually marked using the software Kinovea). Our best resulting markerless body tracker particularly adapted for sprint biomechanics is termed VideoRun2D. The experimental development and assessment of VideoRun2D is reported on forty sprints recorded with a video camera from 5 different subjects, focusing our analysis in 3 key angles in sprint biomechanics: inclination of the trunk, flex extension of the hip and the knee. The CoTracker method showed huge differences compared to the manual labeling approach. However, the angle curves were correctly estimated by the MoveNet method, finding errors between 3.2° and 5.5°. In conclusion, our proposed VideoRun2D based on MoveNet core seems to be a helpful tool for evaluating sprint kinematics in some scenarios. On the other hand, the observed precision of this first version of VideoRun2D as a markerless sprint analysis system may not be yet enough for highly demanding applications. Future research lines towards that purpose are also discussed at the end: better tracking post-processing and user- and time-dependent adaptation.

Laura Pedrouzo-Rodriguez, Luis F. Gomez, Ruben Tolosana et al.

Recent advances in photorealistic avatar generation have enabled highly realistic talking-head avatars, raising security concerns regarding identity impersonation in AI-mediated communication. To advance in this challenging problem, the task of avatar fingerprinting aims to determine whether two avatar videos are driven by the same human operator or not. However, current public databases in the literature are scarce and based solely on old-fashioned talking-head avatar generators, not representing realistic scenarios for the current task of avatar fingerprinting. To overcome this situation, the present article introduces AVAPrintDB, a new publicly available multi-generator talking-head avatar database for avatar fingerprinting. AVAPrintDB is constructed from two audiovisual corpora and three state-of-the-art avatar generators (GAGAvatar, LivePortrait, HunyuanPortrait), representing different synthesis paradigms, and includes both self- and cross-reenactments to simulate legitimate usage and impersonation scenarios. Building on this database, we also define a standardized and reproducible benchmark for avatar fingerprinting, considering public state-of-the-art avatar fingerprinting systems and exploring novel methods based on Foundation Models (DINOv2 and CLIP). Also, we conduct a comprehensive analysis under generator and dataset shift. Our results show that, while identity-related motion cues persist across synthetic avatars, current avatar fingerprinting systems remain highly sensitive to changes in the synthesis pipeline and source domain. The AVAPrintDB, benchmark protocols, and avatar fingerprinting systems are publicly available to facilitate reproducible research.

Pablo Jimenez-Lizcano, Sergio Romero-Tapiador, Ruben Tolosana et al.

Diabetic retinopathy (DR) and diabetic macular edema (DME) are leading causes of preventable blindness among working-age adults. Traditional approaches in the literature focus on standard color fundus photography (CFP) for the detection of these conditions. Nevertheless, recent ultra-widefield imaging (UWF) offers a significantly wider field of view in comparison to CFP. Motivated by this, the present study explores state-of-the-art deep learning (DL) methods and UWF imaging on three clinically relevant tasks: i) image quality assessment for UWF, ii) identification of referable diabetic retinopathy (RDR), and iii) identification of DME. Using the publicly available UWF4DR Challenge dataset, released as part of the MICCAI 2024 conference, we benchmark DL models in the spatial (RGB) and frequency domains, including popular convolutional neural networks (CNNs) as well as recent vision transformers (ViTs) and foundation models. In addition, we explore a final feature-level fusion to increase robustness. Finally, we also analyze the decisions of the DL models using Grad-CAM, increasing the explainability. Our proposal achieves consistently strong performance across all architectures, underscoring the competitiveness of emerging ViTs and foundation models and the promise of feature-level fusion and frequency-domain representations for UWF analysis.

Javier Irigoyen, Roberto Daza, Aythami Morales et al.

This work introduces EduEVAL-DB, a dataset based on teacher roles designed to support the evaluation and training of automatic pedagogical evaluators and AI tutors for instructional explanations. The dataset comprises 854 explanations corresponding to 139 questions from a curated subset of the ScienceQA benchmark, spanning science, language, and social science across K-12 grade levels. For each question, one human-teacher explanation is provided and six are generated by LLM-simulated teacher roles. These roles are inspired by instructional styles and shortcomings observed in real educational practice and are instantiated via prompt engineering. We further propose a pedagogical risk rubric aligned with established educational standards, operationalizing five complementary risk dimensions: factual correctness, explanatory depth and completeness, focus and relevance, student-level appropriateness, and ideological bias. All explanations are annotated with binary risk labels through a semi-automatic process with expert teacher review. Finally, we present preliminary validation experiments to assess the suitability of EduEVAL-DB for evaluation. We benchmark a state-of-the-art education-oriented model (Gemini 2.5 Pro) against a lightweight local Llama 3.1 8B model and examine whether supervised fine-tuning on EduEVAL-DB supports pedagogical risk detection using models deployable on consumer hardware.

Sergio Romero-Tapiador, Ruben Tolosana, Blanca Lacruz-Pleguezuelos et al.

Automatic dietary assessment based on food images remains a challenge, requiring precise food detection, segmentation, and classification. Vision-Language Models (VLMs) offer new possibilities by integrating visual and textual reasoning. In this study, we evaluate six state-of-the-art VLMs (ChatGPT, Gemini, Claude, Moondream, DeepSeek, and LLaVA), analyzing their capabilities in food recognition at different levels. For the experimental framework, we introduce the FoodNExTDB, a unique food image database that contains 9,263 expert-labeled images across 10 categories (e.g., "protein source"), 62 subcategories (e.g., "poultry"), and 9 cooking styles (e.g., "grilled"). In total, FoodNExTDB includes 50k nutritional labels generated by seven experts who manually annotated all images in the database. Also, we propose a novel evaluation metric, Expert-Weighted Recall (EWR), that accounts for the inter-annotator variability. Results show that closed-source models outperform open-source ones, achieving over 90% EWR in recognizing food products in images containing a single product. Despite their potential, current VLMs face challenges in fine-grained food recognition, particularly in distinguishing subtle differences in cooking styles and visually similar food items, which limits their reliability for automatic dietary assessment. The FoodNExTDB database is publicly available at https://github.com/AI4Food/FoodNExtDB.

Javier Hernandez-Ortega, Julian Fierrez, Luis F. Gomez et al.

In this paper we develop FaceQvec, a software component for estimating the conformity of facial images with each of the points contemplated in the ISO/IEC 19794-5, a quality standard that defines general quality guidelines for face images that would make them acceptable or unacceptable for use in official documents such as passports or ID cards. This type of tool for quality assessment can help to improve the accuracy of face recognition, as well as to identify which factors are affecting the quality of a given face image and to take actions to eliminate or reduce those factors, e.g., with postprocessing techniques or re-acquisition of the image. FaceQvec consists of the automation of 25 individual tests related to different points contemplated in the aforementioned standard, as well as other characteristics of the images that have been considered to be related to facial quality. We first include the results of the quality tests evaluated on a development dataset captured under realistic conditions. We used those results to adjust the decision threshold of each test. Then we checked again their accuracy on a evaluation database that contains new face images not seen during development. The evaluation results demonstrate the accuracy of the individual tests for checking compliance with ISO/IEC 19794-5. FaceQvec is available online (https://github.com/uam-biometrics/FaceQvec).

Ivan DeAndres-Tame, Ruben Tolosana, Pietro Melzi et al.

Synthetic data is gaining increasing relevance for training machine learning models. This is mainly motivated due to several factors such as the lack of real data and intra-class variability, time and errors produced in manual labeling, and in some cases privacy concerns, among others. This paper presents an overview of the 2nd edition of the Face Recognition Challenge in the Era of Synthetic Data (FRCSyn) organized at CVPR 2024. FRCSyn aims to investigate the use of synthetic data in face recognition to address current technological limitations, including data privacy concerns, demographic biases, generalization to novel scenarios, and performance constraints in challenging situations such as aging, pose variations, and occlusions. Unlike the 1st edition, in which synthetic data from DCFace and GANDiffFace methods was only allowed to train face recognition systems, in this 2nd edition we propose new sub-tasks that allow participants to explore novel face generative methods. The outcomes of the 2nd FRCSyn Challenge, along with the proposed experimental protocol and benchmarking contribute significantly to the application of synthetic data to face recognition.

Pietro Melzi, Christian Rathgeb, Ruben Tolosana et al.

This study investigates the possibility of mitigating the demographic biases that affect face recognition technologies through the use of synthetic data. Demographic biases have the potential to impact individuals from specific demographic groups, and can be identified by observing disparate performance of face recognition systems across demographic groups. They primarily arise from the unequal representations of demographic groups in the training data. In recent times, synthetic data have emerged as a solution to some problems that affect face recognition systems. In particular, during the generation process it is possible to specify the desired demographic and facial attributes of images, in order to control the demographic distribution of the synthesized dataset, and fairly represent the different demographic groups. We propose to fine-tune with synthetic data existing face recognition systems that present some demographic biases. We use synthetic datasets generated with GANDiffFace, a novel framework able to synthesize datasets for face recognition with controllable demographic distribution and realistic intra-class variations. We consider multiple datasets representing different demographic groups for training and evaluation. Also, we fine-tune different face recognition systems, and evaluate their demographic fairness with different metrics. Our results support the proposed approach and the use of synthetic data to mitigate demographic biases in face recognition.

Ivan DeAndres-Tame, Ruben Tolosana, Pietro Melzi et al.

Synthetic data is gaining increasing popularity for face recognition technologies, mainly due to the privacy concerns and challenges associated with obtaining real data, including diverse scenarios, quality, and demographic groups, among others. It also offers some advantages over real data, such as the large amount of data that can be generated or the ability to customize it to adapt to specific problem-solving needs. To effectively use such data, face recognition models should also be specifically designed to exploit synthetic data to its fullest potential. In order to promote the proposal of novel Generative AI methods and synthetic data, and investigate the application of synthetic data to better train face recognition systems, we introduce the 2nd FRCSyn-onGoing challenge, based on the 2nd Face Recognition Challenge in the Era of Synthetic Data (FRCSyn), originally launched at CVPR 2024. This is an ongoing challenge that provides researchers with an accessible platform to benchmark i) the proposal of novel Generative AI methods and synthetic data, and ii) novel face recognition systems that are specifically proposed to take advantage of synthetic data. We focus on exploring the use of synthetic data both individually and in combination with real data to solve current challenges in face recognition such as demographic bias, domain adaptation, and performance constraints in demanding situations, such as age disparities between training and testing, changes in the pose, or occlusions. Very interesting findings are obtained in this second edition, including a direct comparison with the first one, in which synthetic databases were restricted to DCFace and GANDiffFace.

Daniel DeAlcala, Aythami Morales, Julian Fierrez et al.

This article introduces the Membership Inference Test (MINT), a novel approach that aims to empirically assess if given data was used during the training of AI/ML models. Specifically, we propose two MINT architectures designed to learn the distinct activation patterns that emerge when an Audited Model is exposed to data used during its training process. These architectures are based on Multilayer Perceptrons (MLPs) and Convolutional Neural Networks (CNNs). The experimental framework focuses on the challenging task of Face Recognition, considering three state-of-the-art Face Recognition systems. Experiments are carried out using six publicly available databases, comprising over 22 million face images in total. Different experimental scenarios are considered depending on the context of the AI model to test. Our proposed MINT approach achieves promising results, with up to 90\% accuracy, indicating the potential to recognize if an AI model has been trained with specific data. The proposed MINT approach can serve to enforce privacy and fairness in several AI applications, e.g., revealing if sensitive or private data was used for training or tuning Large Language Models (LLMs).

Daniel DeAlcala, Gonzalo Mancera, Julian Fierrez et al.

We present the Membership Inference Test (MINT) Demo 2, a framework designed to improve transparency in machine learning training processes. MINT is a technique for experimentally determining whether specific data were used during machine learning model training. We establish the theoretical framework and propose multiple architectures for MINT depending on the amount of information known about the models that are being audited. Experimental results using a popular face recognition model, 4 state-of-the-art LLMs, and multiple, diverse, and large-scale public image and text databases achieve promising accuracy levels in the detection of training data of up to 90%. Building on these results, we introduce a comprehensive web platform1 that expands these capabilities to image and text modalities. The platform integrates a diverse technological stack, including MINT, aMINT, and gMINT, allowing users to audit a wide range of models. This demonstrator aims to promote AI transparency and provides a practical tool to foster compliance with emerging AI regulations.

Gonzalo Mancera, Daniel DeAlcala, Aythami Morales et al.

We present LoRA-MINT, a new methodology for Membership Inference Test (MINT) applied to recent Large Language Models (LLMs) fine-tuned for specific Natural Language Processing (NLP) tasks through Low-Rank Adaptation (LoRA). The primary goal is to assess whether individual samples were part of the training data of these adapted models, providing a useful auditing tool for the management of intellectual property and sensitive data. Our analysis explores the relationship between model perplexity and membership status, providing a systematic framework for estimating data exposure in fine-tuned LLMs. We conducted experiments on four models and three benchmark datasets, obtaining precision values in determining if given data were used for training ranging from 0.77 to 0.92, which outperform state-of-the-art baselines and demonstrate the robustness and generality of the proposed method. In general, our findings underscore the potential of LoRA-MINT as an effective and scalable framework for auditing LLMs, improving transparency, and fostering the ethical and responsible deployment of AI and NLP technologies. For the sake of concreteness and current relevance, our discussion and experiments are centered on LoRAadjusted LLMs, but note that most of the presented methodology is easily applicable for auditing training data given any other technique for adapting LLMs or, more generally, any other domain-adapted AI models.

Álvaro Becerra, Javier Irigoyen, Roberto Daza et al.

In this article, we explore computer vision approaches to detect abnormal head pose during e-learning sessions and we introduce a study on the effects of mobile phone usage during these sessions. We utilize behavioral data collected from 120 learners monitored while participating in a MOOC learning sessions. Our study focuses on the influence of phone-usage events on behavior and physiological responses, specifically attention, heart rate, and meditation, before, during, and after phone usage. Additionally, we propose an approach for estimating head pose events using images taken by the webcam during the MOOC learning sessions to detect phone-usage events. Our hypothesis suggests that head posture undergoes significant changes when learners interact with a mobile phone, contrasting with the typical behavior seen when learners face a computer during e-learning sessions. We propose an approach designed to detect deviations in head posture from the average observed during a learner's session, operating as a semi-supervised method. This system flags events indicating alterations in head posture for subsequent human review and selection of mobile phone usage occurrences with a sensitivity over 90%.

Roberto Daza, Alvaro Becerra, Ruth Cobos et al.

This work presents the IMPROVE dataset, a multimodal resource designed to evaluate the effects of mobile phone usage on learners during online education. It includes behavioral, biometric, physiological, and academic performance data collected from 120 learners divided into three groups with different levels of phone interaction, enabling the analysis of the impact of mobile phone usage and related phenomena such as nomophobia. A setup involving 16 synchronized sensors-including EEG, eye tracking, video cameras, smartwatches, and keystroke dynamics-was used to monitor learner activity during 30-minute sessions involving educational videos, document reading, and multiple-choice tests. Mobile phone usage events, including both controlled interventions and uncontrolled interactions, were labeled by supervisors and refined through a semi-supervised re-labeling process. Technical validation confirmed signal quality, and statistical analyses revealed biometric changes associated with phone usage. The dataset is publicly available for research through GitHub and Science Data Bank, with synchronized recordings from three platforms (edBB, edX, and LOGGE), provided in standard formats (.csv, .mp4, .wav, and .tsv), and accompanied by a detailed guide.

Roberto Daza, Javier Irigoyen, Ivan Lopez et al.

This work introduces ARES, a platform and open pilot dataset for auditing adaptive social engineering risks in LLM-mediated social decision-making through controlled social games. ARES supports human--human, human--AI, and AI--AI settings, combining configurable game templates, role-conditioned LLM agents, psychology-informed participant profiling, structured interaction trees, and synchronised behavioural and biometric acquisition, filtering, and deep-learning-based feature extraction. The pilot dataset was collected from 15 participants interacting with a role-conditioned GPT-5.4 agent in two concatenated games: an adapted Prisoner's Dilemma and an Ultimatum Game. It comprises 340 GB of raw and processed multimodal data across six streams: interaction logs, video, screen recordings, gaze logs, smartwatch signals, and game/questionnaire metadata. These data include interaction paths, written justifications, psychological profiles, subjective feedback, perceived counterpart identity, game outcomes, and derived behavioural, facial, and gaze features. Alongside the dataset, we provide descriptive analyses to characterise the pilot release. Rigorous risk evaluation is essential for the deployment of secure AI systems, as it enables the identification and mitigation of vulnerabilities, ensures the protection of sensitive data, and supports compliance with evolving regulatory and ethical standards in society.

Alejandro Peña, Julian Fierrez, Aythami Morales et al.

The use of language technologies in high-stake settings is increasing in recent years, mostly motivated by the success of Large Language Models (LLMs). However, despite the great performance of LLMs, they are are susceptible to ethical concerns, such as demographic biases, accountability, or privacy. This work seeks to analyze the capacity of Transformers-based systems to learn demographic biases present in the data, using a case study on AI-based automated recruitment. We propose a privacy-enhancing framework to reduce gender information from the learning pipeline as a way to mitigate biased behaviors in the final tools. Our experiments analyze the influence of data biases on systems built on two different LLMs, and how the proposed framework effectively prevents trained systems from reproducing the bias in the data.

Roberto Daza, Lin Shengkai, Aythami Morales et al.

This work introduces SMARTe-VR, a platform for student monitoring in an immersive virtual reality environment designed for online education. SMARTe-VR aims to collect data for adaptive learning, focusing on facial biometrics and learning metadata. The platform allows instructors to create customized learning sessions with video lectures, featuring an interface with an AutoQA system to evaluate understanding, interaction tools (for example, textbook highlighting and lecture tagging), and real-time feedback. Furthermore, we released a dataset that contains 5 research challenges with data from 10 users in VR-based TOEIC sessions. This data set, which spans more than 25 hours, includes facial features, learning metadata, 450 responses, difficulty levels of the questions, concept tags, and understanding labels. Alongside the database, we present preliminary experiments using Item Response Theory models, adapted for understanding detection using facial features. Two architectures were explored: a Temporal Convolutional Network for local features and a Multilayer Perceptron for global features.

Giuseppe Stragapede, Ruben Vera-Rodriguez, Ruben Tolosana et al.

This paper describes the results of the IEEE BigData 2023 Keystroke Verification Challenge (KVC), that considers the biometric verification performance of Keystroke Dynamics (KD), captured as tweet-long sequences of variable transcript text from over 185,000 subjects. The data are obtained from two of the largest public databases of KD up to date, the Aalto Desktop and Mobile Keystroke Databases, guaranteeing a minimum amount of data per subject, age and gender annotations, absence of corrupted data, and avoiding excessively unbalanced subject distributions with respect to the considered demographic attributes. Several neural architectures were proposed by the participants, leading to global Equal Error Rates (EERs) as low as 3.33% and 3.61% achieved by the best team respectively in the desktop and mobile scenario, outperforming the current state of the art biometric verification performance for KD. Hosted on CodaLab, the KVC will be made ongoing to represent a useful tool for the research community to compare different approaches under the same experimental conditions and to deepen the knowledge of the field.

Gonzalo Mancera, Aythami Morales, Julian Fierrez et al.

The use of Natural Language Processing (NLP) in highstakes AI-based applications has increased significantly in recent years, especially since the emergence of Large Language Models (LLMs). However, despite their strong performance, LLMs introduce important legal/ ethical concerns, particularly regarding privacy, data protection, and transparency. Due to these concerns, this work explores the use of Named- Entity Recognition (NER) to facilitate the privacy-preserving training (or adaptation) of LLMs. We propose a framework that uses NER technologies to anonymize sensitive information in text data, such as personal identities or geographic locations. An evaluation of the proposed privacy-preserving learning framework was conducted to measure its impact on user privacy and system performance in a particular high-stakes and sensitive setup: AI-based resume scoring for recruitment processes. The study involved two language models (BERT and RoBERTa) and six anonymization algorithms (based on Presidio, FLAIR, BERT, and different versions of GPT) applied to a database of 24,000 candidate profiles. The findings indicate that the proposed privacy preservation techniques effectively maintain system performance while playing a critical role in safeguarding candidate confidentiality, thus promoting trust in the experimented scenario. On top of the proposed privacy-preserving approach, we also experiment applying an existing approach that reduces the gender bias in LLMs, thus finally obtaining our proposed Privacyand Bias-aware LLMs (PBa-LLMs). Note that the proposed PBa-LLMs have been evaluated in a particular setup (resume scoring), but are generally applicable to any other LLM-based AI application.

Daniel DeAlcala, Aythami Morales, Julian Fierrez et al.

We present the Membership Inference Test Demonstrator, to emphasize the need for more transparent machine learning training processes. MINT is a technique for experimentally determining whether certain data has been used during the training of machine learning models. We conduct experiments with popular face recognition models and 5 public databases containing over 22M images. Promising results, up to 89% accuracy are achieved, suggesting that it is possible to recognize if an AI model has been trained with specific data. Finally, we present a MINT platform as demonstrator of this technology aimed to promote transparency in AI training.

Miguel Lopez-Duran, Julian Fierrez, Aythami Morales et al.

The automatic analysis of document layouts in digital-born PDF documents remains a challenging problem due to the heterogeneous arrangement of textual and nontextual elements and the imprecision of the textual metadata in the Portable Document Format. In this work, we benchmark Graph Neural Network (GNN) architectures for the task of fine-grained layout classification of text blocks from digital native documents. We introduce two graph construction structures: a k-closest-neighbor graph and a fully connected graph, and generate node features via pre-trained text and vision models, thus avoiding manual feature engineering. Three experimental frameworks are evaluated: single-modality (text or visual), concatenated multimodal, and dual-branch multimodal. We evaluated four foundational GNN models and compared them with the baseline. Our experiments are specifically conducted on a rich dataset of public affairs documents that includes more than 20 sources (e.g., regional and national-level official gazettes), 37K PDF documents, with 441K pages in total. Our results demonstrate that GraphSAGE operating on the k-closest-neighbor graph in a dual-branch configuration achieves the highest per-class and overall accuracy, outperforming the baseline in some sources. These findings confirm the importance of local layout relationships and multimodal fusion exploited through GNNs for the analysis of native digital document layouts.

Eric Macias-Fassio, Aythami Morales, Cristina Pruenza et al.

The biomedical field is among the sectors most impacted by the increasing regulation of Artificial Intelligence (AI) and data protection legislation, given the sensitivity of patient information. However, the rise of synthetic data generation methods offers a promising opportunity for data-driven technologies. In this study, we propose a statistical approach for synthetic data generation applicable in classification problems. We assess the utility and privacy implications of synthetic data generated by Kernel Density Estimator and K-Nearest Neighbors sampling (KDE-KNN) within a real-world context, specifically focusing on its application in sepsis detection. The detection of sepsis is a critical challenge in clinical practice due to its rapid progression and potentially life-threatening consequences. Moreover, we emphasize the benefits of KDE-KNN compared to current synthetic data generation methodologies. Additionally, our study examines the effects of incorporating synthetic data into model training procedures. This investigation provides valuable insights into the effectiveness of synthetic data generation techniques in mitigating regulatory constraints within the biomedical field.

Daniel DeAlcala, Aythami Morales, Julian Fierrez et al.

Active Membership Inference Test (aMINT) is a method designed to detect whether given data were used during the training of machine learning models. In Active MINT, we propose a novel multitask learning process that involves training simultaneously two models: the original or Audited Model, and a secondary model, referred to as the MINT Model, responsible for identifying the data used for training the Audited Model. This novel multi-task learning approach has been designed to incorporate the auditability of the model as an optimization objective during the training process of neural networks. The proposed approach incorporates intermediate activation maps as inputs to the MINT layers, which are trained to enhance the detection of training data. We present results using a wide range of neural networks, from lighter architectures such as MobileNet to more complex ones such as Vision Transformers, evaluated in 5 public benchmarks. Our proposed Active MINT achieves over 80% accuracy in detecting if given data was used for training, significantly outperforming previous approaches in the literature. Our aMINT and related methodological developments contribute to increasing transparency in AI models, facilitating stronger safeguards in AI deployments to achieve proper security, privacy, and copyright protection.