1,720,958 research outputs found
Supervised and Semi-Supervised Explainable AI applied to Cardiovascular Magnetic Resonance Mapping Techniques
No full textDeep Learning techniques have demonstrated broad applicability across numerous domains, including medicine. Among these, Cardiovascular Magnetic Resonance Imaging stands out as a field where Deep Learning has the potential to transform myocardial tissue characterization through the automated, objective, and reproducible assessment of modern-day quantitative techniques like T1 and T2 mapping images, which serve as the case study of this Thesis. The clinical implementation of these techniques, however, is hindered by several obstacles, including the need for center-specific reference values -- which strongly depend on the magnetic field strength, acquisition sequence, and chosen parameters -- along with the scarcity of large annotated datasets, class imbalance, and the requirement for transparent and explainable decision-making. This Thesis proposes a Deep Learning approach that utilizes supervised and semi-supervised learning and model ensembling to navigate some of these challenges. It provides explainability tools to support transparent case-level explanation of model predictions, facilitate clinician trust and supporting clinical adoption. Experimental results demonstrate that the proposed framework enhanced predictive accuracy and reliability harnessing actionable insights in Cardiac Imaging
A Survey of Modern Hybrid Particle Swarm Optimization Algorithms
No full textBio-inspired, population-based meta-heuristic for global optimization are very popular algorithms for addressing complex computational problems that traditional methods struggle to solve. Among the existing algorithms, the swarm intelligence algorithm Particle Swarm Optimization (PSO) is one of the most popular, thanks to its simplicity and effectiveness in multiple scenarios. This article focuses on recent hybrid optimization methods that extend the basic functioning of PSO. Hybridization, in this context, is defined as the integration of PSO with a different technique, to take advantage of the strengths of both algorithms. According to our findings, many variants have been proposed. The most frequent solutions consist of the hybridization of PSO with evolutionary operators (e.g. Genetic Algorithms and Differential Evolution); such strategies usually maintain a high degree of diversity into the population, enhancing global search capability, while reducing the risk of stagnation. Meanwhile the most widespread applications are from the areas of energy optimization, structural engineering and machine learning problems, demonstrating the versatility of these hybrid approaches
Assessing Cardiac Functionality by Means of Interpretable AI and Myocardial Strain
No full textCardiac Imaging is a powerful methodology for the accurate assessment of heart functionality. Among the possible approaches, Myocardial Strain assesses the functionality of the heart by tracking the movement and deformation of myocardium during the cardiac cycle. This information, that can be acquired also by means of Cardiac Magnetic Resonance, can pave the way to the development of predictive models using machine learning. In this work, we developed a predictive model of left ventricular ejection fraction, which is a measure of the heart’s function to pump oxygen-rich blood to the body, trained using strain data. Specifically, we developed a fully interpretable model based on a rule-based Fuzzy Inference System, coupled with a novel methodology for the disambiguation of the rules. Our results show that the developed model is able to accurately estimate the ejection fraction, and can provide physicians with additional insights about the role of strain features
Improving the Efficiency and the Validity of Molecular Transformers
No full textSince their advent, Transformer models have been applied across a wide range of fields, including cheminformatics. In this context, drug discovery has benefited from using Molecular Transformers by leveraging diverse string representations of molecules, such as the Simplified Molecular Input Line Entry Systems (SMILES), for a variety of tasks. In this study, we present a model focused on the optimization of a formerly developed Molecular Transformer specifically dedicated to metabolism prediction. Metabolism refers to all the biotransformations a drug undergoes once inside the human body, directly influencing its therapeutic effect and potential toxicity, and therefore represents a key topic in medicinal chemistry. Framing molecular transformation prediction as a sequence-to-sequence translation task has shown promise, but suffers from limitations such as low validity of generated molecules and high computational cost. To address this limitation, we here propose an optimized model that integrates pre-training, transfer learning, and fine-tuning techniques, already improving validity and reducing computation time. Finally, by separating the metabolism prediction task from the SMILES syntax learning, we ensure broader applicability of the proposed model across diverse datasets and a variety of SMILES-based tasks beyond metabolic transformations, expanding its potential utility
Going Beyond Counting First Authors in Author Co-citation Analysis
Get PDFThe present study examines one of the fundamental aspects of author co-citation analysis (ACA) - the way co-citation
counts are defined. Co-citation counting provides the data on which all subsequent statistical analyses and mappings
are based, and we compare ACA results based on two different types of co-citation counting - the traditional type that
only counts the first one among a cited work's authors on the one hand and a non-traditional type that takes into
account the first 5 authors of a cited work on the other hand. Results indicate that the picture produced through this non-traditional author co-citation counting contains more coherent author groups and is therefore considerably clearer. However, this picture represents fewer specialties in the research field being studied than that produced through the traditional first-author co-citation counting when the same number of top-ranked authors is selected and analyzed. Reasons for these effects are discussed
Variations on the Author
Get PDF“Variations on the Author” discusses two of Eduardo Coutinho’s recent films (Um Dia na Vida, from 2010, and Últimas Conversas, posthumously released in 2015) and their contribution to the general question of documentary authorship. The director’s filmography is characterized by a consistent yet self-effacing form of authorial self-inscription: Coutinho often features as an interviewer that rather than express opinions propels discourses; an interviewer that is good at listening. This mode of self-inscription characterizes him as an author who is not expressive but who is nonetheless markedly present on the screen. In Um Dia na Vida, however, Coutinho is completely absent form the image, while Últimas Conversas, on the contrary, includes a confessional prologue that moves the director from the margins to the center of his films. This article examines the ways in which these works stand out in the filmography of a director who offers new insights into the notion of cinematic authorship
Appropriate Similarity Measures for Author Cocitation Analysis
Get PDFWe provide a number of new insights into the methodological discussion about author cocitation analysis. We first argue that the use of the Pearson correlation for measuring the similarity between authors’ cocitation profiles is not very satisfactory. We then discuss what kind of similarity measures may be used as an alternative to the Pearson correlation. We consider three similarity measures in particular. One is the well-known cosine. The other two similarity measures have not been used before in the bibliometric literature. Finally, we show by means of an example that our findings have a high practical relevance.information science;Pearson correlation;cosine;similarity measure;author cocitation analysis
Massive stochastic simulation of cosmic rays propagation in the heliosphere: The COSMICA code
No full textThe accurate modeling of galactic cosmic ray (GCR) propagation in the heliosphere requires solving the Parker Transport Equation (PTE), a multidimensional nonlinear equation that cannot be addressed analytically without strong approximations. In recent decades, stochastic differential equation (SDE)–Monte Carlo methods have emerged as a powerful numerical strategy for this problem, thanks to their numerical stability, relatively low memory requirements, and intrinsic parallelism. The increasing availability of general-purpose Graphics Processing Units (GPUs) has further revolutionized this approach by enabling massive parallelization of particle trajectories at relatively low cost. In this work, we introduce COSMICA (COde for a Speedy Montecarlo Involving Cuda Architecture), a new open-source multi-GPU code written in CUDA/C++ for the three-dimensional solution of the PTE. COSMICA has been specifically designed to optimize GPU resource usage and scalability, with strategies including memory hierarchy exploitation, register-conscious kernel design, warp-aware scheduling, and parameter reordering for multi-GPU execution. Benchmark results demonstrate that COSMICA reduces runtimes from weeks to hours for large-scale simulations. These optimizations make COSMICA a versatile tool for systematic studies of cosmic-ray modulation and parameter exploration, thereby expanding the feasibility of investigations that were previously computationally prohibitive. The present article constitutes the first part of a two-paper series, focusing on code design and computational performance; a companion paper will present its validation against benchmark models
COSMICA: A GPU-Optimized Code for Solar Modulation Studies
No full textWe present COSMICA, an opensource high-performance GPU-accelerated numerical code for modeling cosmic ray solar modulation, and its application to study CR diffusion parameters. Developed within the framework of the ICSC-Italian Research Center on High-Performance Computing, Big Data and Quantum Computing (Spoke-3), COSMICA is undergoing continuous software optimization to maximize efficiency on NVIDIA architectures. COSMICA is coupled with SDEGnO, another ICSC project, designed for the efficient parameter tuning, exploring the large parameter space in solar modulation studies. As a first physical use-case study, we exploit COSMICA to investigate Forbush Decreases (FDs), which are transient cosmic ray intensity reductions caused by interplanetary disturbances. The analysis leverages the high-precision daily measurements from AMS-02, which provide cosmic ray fluxes across a wide range of rigidities. The ability to simultaneously study not only protons but also helium isotopes offers complementary insights into charge- and mass-dependent transport effects. By analyzing FD events, we assess localized variations in diffusion parameters and their impact on cosmic ray transport. The results confirm the stability of the rigidity dependence of the diffusion tensor, supporting the use of FDs as probes of localized well constrained heliospheric conditions. The computational efficiency of COSMICA paves the way for large-scale simulations, systematic FD catalogue analysis and a more in-depth understanding of the parameter that regulates the solar modulation, together with their dependencies
- …
