In this paper we investigate the discrete implication functions that satisfy the Modus Ponens with respect to a discrete uninorm. From this study, and analogously to the case of implication functions defined in the unit interval, it is obtained that the uninorms to be considered must be conjunctive. For this reason, we want to study the case of discrete implications derived from discrete uninorms. In particular, we investigate the U-conditionality for discrete residual implications derived from discrete uninorms in U-min and discrete idempotent uninorms.
As the usage of the edge-cloud continuum rises, Kubernetes presents itself as a solution that allows easy control and deployment of applications in these highly-distributed and heterogeneous environments. In this context, Artificial Intelligence methods have been proposed to aid in the task allocation process to optimize different aspects of the system, such as application execution time, load balancing or energy consumption. In this paper, we propose a space-time combinational model that uses Deep Reinforcement Learning (DRL) to recommend node allocations for Kubernetes pods with the objective of optimizing the overall energy consumption of the cluster while maintaining pod execution ratio. In particular, our approach uses Proximal Policy Optimization (PPO) with custom Neural Networks to train a DRL agent and includes a custom Kubernetes operator to enforce allocations based on the node recommendations generated by the agent. Using our custom solution, we performed a series of experiments with different workloads and compared the performance with the base Kubernetes scheduler. Our experimental results demonstrate a notable reduction of up to 24% in the energy consumption of the Kubernetes cluster.
Multiple sclerosis (MS) is a chronic autoimmune disease that predominantly affects the central nervous system (CNS) and is a leading cause of neurological disability among young adults. MS diagnosis heavily relies on clinical symptoms coupled with the detection of demyelinating lesions in the CNS, as depicted in conventional magnetic resonance images (MRI). These lesions may evolve over time and serve as critical biomarkers for assessing disease activity and the effectiveness of therapeutic interventions. Manual lesion delimitation in images is a tedious and error-prone process, prompting active research into automated systems. MS lesion segmentation can be conducted in two very different approaches: cross-sectionally: by using a single MRI, or longitudinally, by identifying changes such as new lesions or variations in lesion size across consecutives MRIs. Deep learning architectures such as U-Net, which employs convolutional neural networks with skip connections, have proven effective for MRI lesion segmentation. In this paper, we utilize the nnUNet v2 architecture and introduce an enhanced method capable of segmenting both cross-sectional and longitudinal MRIs. We used the ISBI 2015 dataset, training our U-Net model on a subset of patients using a generative pipeline and testing on another subset of patients from the same dataset. For training the model, we used a Bayesian generative approach to create different synthetic data simulating random variations in lesion masks. In this process, we have randomly simulated multiple temporal variations of the lesion masks, including erosion, dilation, and the removal of individual lesions within each mask. Then, we adapted the U-Net to train using a single MRI and its corresponding synthetic segmentation, which simulates the lesions in a previous time point. Thanks to the large amount of synthetic data generated, the U-Net is able to learn the intrinsic behavior of the lesions which include expansion, shrinkage, appearance and disappearance. This approach enables, for the first time, the two classical operational modes for MS lesion segmentation within a single model: cross-sectional segmentation, where the model receives a MRI and an empty mask as input, and longitudinal segmentation, where the input includes a MRI and a mask from a previous time point. Our approach achieved a Dice Similarity Coefficient (DSC) of 0.75 for cross-sectional segmentation and 0.81 for longitudinal segmentation, indicating improved performance when incorporating temporal information.This study demonstrates that it is possible to apply a generative method in conjunction with a U-Net architecture for both cross-sectional and longitudinal MRI lesion segmentation in MS, yielding promising results.
Stress, the body’s reaction to threatening situation, may derive in negative physical and mental health effects. Since speech under stress (SuS) shows relevant changes in the voice production system, its analysis may become instrumental to identify stressful situations non-invasively. Despite most investigations have focused on identifying SuS from prosodic and spectral variations, preliminary attempts have evaluated the alterations of the glottal source. This work evaluates the viability of considering glottal source features for SuS classification. Results on the cardinal vowels extracted from VOCE corpus show classification accuracies between 90.5% and 92.3% when glottal source features are combined with Mel Frequency Cepstral Coefficients to identify SuS, improving results in 6 out of 8 the SVM and NN two-class classifiers when trained with them separately.
Breast density is a crucial biomarker for predicting BC risk and recurrence. Women with dense breast tissues have a higher likelihood of developing BC, and dense tissue can obscure lesions, reducing detection sensitivity. Mammograms are vital for evaluating breast density, typically classified using the BI-RADS system. The main challenge in breast density segmentation is accurately localizing dense tissues. While segmentation models require detailed pixel-wise annotations, obtaining these labels is time-consuming and requires medical expertise. This paper proposes a weakly supervised approach for breast density localization, allowing deep neural network classifiers to generate saliency maps that highlight dense tissue regions based on image-level labels. We validate this model on the RSNA dataset and achieve a Dice score of 0.754, comparable to state-of-the-art supervised methods.
We formalise human teamwork in tasks involving judgment as a public goods game. Our focus is on tasks where members’ contributions are combined through weighted averaging, such as brain-storming. Using a multiagent system, we examine the alignment between learned agent strategies and Nash Equilibria. Overall, our results demonstrate that our multiagent system effectively approximates the Nash Equilibria of the game.
This study delves into the characterization of synthetic lung nodules using latent diffusion models applied to chest CT scans. Our experiments involve guiding the diffusion process by means of a binary mask for localization and various nodule attributes. In particular, the mask indicates the approximate position of the nodule in the shape of a bounding box, while the other scalar attributes are encoded in an embedding vector. The diffusion model operates in 2D, producing a single synthetic CT slice during inference. The architecture comprises a VQ-VAE encoder to convert between the image and latent spaces, and a U-Net responsible for the denoising process. Our primary objective is to assess the quality of synthesized images as a function of the conditional attributes. We discuss possible biases and whether the model adequately positions and characterizes synthetic nodules. Our findings on the capabilities and limitations of the proposed approach may be of interest for downstream tasks involving limited datasets with non-uniform observations, as it is often the case for medical imaging.
Food is an important factor when choosing a touristic destination, and culinary images are a fundamental tool in gastronomic marketing. This paper presents an approach to analyse the use of color in food images based on a dataset of more than 22000 recipes coming form a popular recipe website, including images of the final dishes and scores that indicate how well the recipe was valued by users. First, the specific cuisine of each recipe is inferred by using the Llama 3 Large Language Model (LLM). The main color of the dish images is also determined. Finally, statistically significant relationships between scores and cuisine labels is also determined. Our preliminary results show some relevant results for tourist marketing, such as showing the most popular cuisines.
As shape alterations in three-dimensional biological structures are associated to numerous pathological processes, quantitative shape analysis for obtaining phenotypic biomarkers of diagnostic potential has become a prominent research area. In this context, the automatic detection of landmarks on 3D anatomical structures is crucial for developing high-throughput phenotyping tools. This study evaluates the performance of multi-view consensus convolutional networks - originally developed for facial landmarking- in automatically detecting landmarks on three different 3D anatomical structures: the face, the upper respiratory airways and the brain hippocampi. Leveraging magnetic resonance imaging datasets, we trained multiple models and assessed their accuracy against manual annotations, while analyzing the impact of different network hyperparameters on the results.
This work in progress aims to help achieve an early diagnosis of cognitive and functional deterioration in elderly people who live alone through the use of Acoustic Event Detection (AED), Wireless Acoustic Sensor Networks (WASN) and Internet of Things (IoT). As of now, four fully functional networks have been deployed and are acquiring data in order to build a custom dataset.
In domains requiring intelligent agents to emulate plausible human-like behavior, such as formative simulations, traditional techniques like behavior trees encounter significant challenges. Large Language Models (LLMs), despite not always yielding optimal solutions, usually offer plausible and human-like responses to a given problem. In this paper, we exploit this capability and propose a novel architecture that integrates an LLM for decision-making with a classical automated planner that can generate sound plans for that decision. The combination aims to equip an agent with the ability to make decisions in various situations, even if they were not anticipated during the design phase.
In this article, we introduce a tool, Popinns, to implement Deep Neural Networks (DNNs) on fixed point architectures. Popinns takes as input the Tensorflow model of a DNN whose coefficients are floating-point numbers and generates a C code in fixed-point arithmetic. The approach implemented in Popinns is based on a formal semantics describing the propagation of the errors through the computations performed by the network. From this semantics, we deduce a system of constraints made of inequalities between linear expressions among integers and of min and max operations. The solution of this system, computed by an optimizing SMT solver, gives the optimal formats of the fixed-point numbers at each point of the DNN. As a result, we synthesize a fixed-point C code that satisfies an error bound set by the user with respect to the initial Tensorflow model. The present article describes Popinns architecture, its features as well as the intermediary and final results computed by the tool.
This work delves into the intricate landscape of employing large language models for the transcription of Catalan texts within the realm of artificial intelligence. As it has been recently stated, Catalan is a medium language that has a very low presence in the digital frameworks and there is still space for improvement in the generation and maturity of the computational linguistics resources for the Catalan language [4]. Nowadays, in the middle of a disruptive digital transformation, the presence of all cultures and languages in Internet and digital frameworks becomes crucial. Generating digital contents in Catalan becomes a priority. Prioritizing the generation of digital contents in Catalan and being able to use Natural Language Processing or Speech recognition will enlarge the impact to other audiences. A central activity is transcription of Catalan videos into Catalan subtitles that eventually can be translated in other languages. Up to now, channels as popular as YouTube still cannot provide a transcription of a Catalan video to generate subtitles in Catalan, and this precludes the possibility to get the subtitles translated to other languages. In Catalonia, some specific policies promote the innovation and advances in the Catalan computational resources, like the [12], or the AINA project [8]. In this work, we analyse the state of the art on transcribing Catalan videos to generate subtitles in three languages (Catalan, Spanish and English). Eventually, the corpus analysed had some Spanish video that has also been considered. The paper rigorously tests a few resources to discern their efficacy in accurately transcribing Catalan text. Despite the promising capabilities of these models, our findings revealed a worse performance than expected, as these systems are still very sensitive to the characteristics of the speaker voice and speech. Common challenges included difficulties in handling Catalan-specific diacritics, idiosyncratic vocabulary, and nuances of regional dialects are identified. The paper describes the experimental setting where several tools have been tested and the results, providing some conclusions and diagnosis of the limitations and strengthnesses of the tested tools.
When a Knowledge Discovery from Data (KDD) (Fayyad, Piatetsky-Shapiro, & Smyth, 1996) process is being applied to get knowledge, several methods could be used (Gibert, et al., 2018). A simple and fast way to obtain preliminary insights from data before using KDD models is by generating a basic descriptive analysis. It is one of the most popular ways to describe experimental data and should be the beginning of all data projects. Nevertheless some of the main knowledge that can be extracted in a descriptive analysis is hidden due to underlying multivariate structures which could be elicited through multivariate analysis techniques. Moreover, the domain expert is key for a proper interpretation of descriptive results. At the same time, there is a lack of automatic reporting techniques that can report and help in the interpretation of complex patterns and the use of advanced multivariate techniques. This paper shows the tool developed to generate automatic interpretation of Multiple Correspondence Analysis (MCA) and Principal Components Analysis (PCA) by using RMarkdown. This tool generates a Word document which contains the automatic interpretation of the results, built on the basis of regular expressions ellaborating over the R analytical outputs (either numerical or graphical results). The proposal is being applied with some real data, like INSESS database on social vulnerabilities of the Catalan population. In conclusion, the developed tool contributes to facilitate the factorial methods results, avoiding the misinterpretation of the results and the involuntary skipping of conclusions due to the large amount of knowledge that can be extracted from a complete factorial analysis. Also, this software enables non-expert users to read multivariate analysis results in a friendly way. Moreover, this tool saves time in the interpretation step and is a basis to support the expert to start the report with the results, even the output of the software could become the report or an intermediate report.
During the last two decades, the operation of the electrical grid has undergone significant changes. This evolution is closely tied to the integration of power electronics into distributed generation systems, which led to increased utilization of renewable energy and as a consequence mitigating climate change by lowering emissions. On the other hand, artificial intelligence plays a considerable role in shaping the development and progress of various technologies, such as the electrical grid. This work presents a study for the application of Reinforcement Learning (RL) tools in distributed generation systems. The objective of using RL is to address voltage perturbations in real time. RL will be useful to maximize the resilience of the system in the event of short circuits of a short duration and minimize the risk of disconnect.
The model of quality of life (QOL) used in Catalonia to assess the QOL level of people with intellectual disability considers eight dimensions defined operationally by core indicators. In this paper, to advance the knowledge about these dimensions regarding social service Catalonia users, we employ logic explained networks to generate global explanations of the correlations between these dimensions and present the results obtained.
Users can share their opinion visiting a restaurant or a hotel by Online Generated Reviews OGRs on platforms such as TripAdvisor Booking or Yelp. Put all together they are thousands of sentences which are quite difficult to seize for a human and to get a comprehensive opinion of the location. This study proposes a Decision Support System DSS composed of three modules extraction of information from TripAdvisor comments summarizing and rating. Compared to prior Research our solution proposes a Neural Network Transformer-based system to summarize and rate thousands of TripAdvisor comments. Our results are bifold. First the analysis of massive comments downloads reveals a bias between the real customer experience based on verbal opinions and the ratings scored in stars. Second we present and online host a DSS which provides a summary of customer experiences per hotel. For Research in Tourism and Hospitality it represents a new milestone in the artificial Intelligence journey and an application of Generative Pretrained Transformer GPT model. For operation Managers it is a novel application of the use of artificial intelligence to embrace the digital revolution. Indeed it helps to determine what customers value most and determine adequate action plan to business requirements.
LHCb is one of the four largest high-energy physics experiments at CERN focused in high precision measurements of particle physics. The LHCb detector has undergone a recent upgrade [1] implying changes at subdetectors, data taking conditions and data processing model. Information from subdetectors is processed at 30MHz at a first trigger phase builded entirely with GPUs to reduce this rate down to 1MHz. Afterwards, the same information is processed in a second trigger phase that runs in CPUs, performing a complete reconstruction and identification of particles. This upgrade implies an evolution of the algorithms used at trigger level. In order to keep performance and speed up processing time, some of them have been replaced by machine learning algorithms. To perform particle identification, one of the LHCb approaches uses a neural network using the information from all subdetectors. In this paper we explain the advantages of this method and the capabilities that machine learning brings to LHCb focused in the global particle identification and throughput improvement achieved with it.
Structural defects, such as cracks, are crucial in various infrastructures, with their accurate delineation paramount for maintenance. However, existing methods often struggle to precisely segment cracks. Despite the advent of deep learning in image segmentation, the recurrent convolution and pooling operations tend to overlook vital edge information, thus compromising the final segmentation accuracy. This paper proposes a pixel-level crack segmentation network using a UNet architecture with a pre-trained ConvNext as the encoder, combined with Multiple Dimension Attention Enhancement (MDAE) blocks. The MDAE block enhances local edge information acquisition, leading to more precise crack segmentation. Experimental results on a public dataset, Crack500, demonstrate the proposed network’s effectiveness, achieving an IoU of 59.6% and an F1-score of 74.7%, thus significantly improving crack segmentation performance.
Producing grapes for wine is a complex and delicate process that requires meticulous execution of numerous tasks to reach the desired quality and yield. Wine producers dedicate many resources to specialized teams, from vine pruning in the early months of the year to harvesting when the fruit is ripe. Scheduling the available workforce and resources to complete these tasks optimally is of utmost importance. In this context, parting from an initial schedule could assist in the decision-making process and enhance the outcomes. With this objective, the Vitigeoss Business Service has been developed within the context of the Vitigeoss project. It is a generalized scheduling tool based on Constraint Programming, which assigns tasks to work teams on specific timeslots. For this, it uses First Fit Decreasing and Local Search as search algorithms. It also offers the capabilities of choosing between a set of constraints defined alongside wine producers, and focusing on one of three objective functions: time, cost, and environmental impact. In this document, a series of updates on the constraint decision process and customization of the domain are introduced, and its implications for the results obtained are analized. The Vitigeoss Business Service was available during the 2023 campaign, and its results have been validated with actual planning data from three use cases, showing improvements in the use of resources and finalization time.
