Programa de Doctorat en Intel·ligència Artificial: Enviaments recents
Ara mostrant els elements 1-20 de 75
Human-on-the-loop continual learning: data, knowledge and agents for model adaptation
(English) How do we effectively adapt a Machine Learning (ML) model when real-world data changes? This seemingly straightforward question underlies one of the most persistent challenges in deploying ML systems in production ...
Supervised and reinforcement learning for extreme adaptive optics with application to exoplanet imaging
(English) This thesis demonstrates that applying a combination of deep learning (DL) models in adaptive optics (AO), each addressing an AO error source, can significantly increase performance, providing telescopes with ...
Exploring the dynamics of the β2-adrenergic receptor: insights from explainable AI in GPCR research
(English) G-protein coupled receptors are transmembrane proteins that serve as critical mediators between extracellular signals and intracellular responses. These highly dynamic entities orchestrate a wide array of cellular ...
Data integration of longitudinal single-cell with multi-omics data to enable precision medicine for complex diseases
(English) Recent developments in high-throughput analysis technologies have transformed biomedical research by generating extensive longitudinal and heterogeneous multi-omics data, including transcriptomics, proteomics, ...
Modeling and reconstruction of 3D humans under context
(English) The study of human's and their behavior through the analysis of images and videos has long been a central topic in Computer Vision. The reconstruction and modeling of human behavior have garnered increasing ...
Advances towards a robust AI-based glioma grading system
(English) Gliomas represent the most common and aggressive form of primary brain tumors in adults, posing significant diagnostic challenges due to their heterogeneous molecular, histological, and radiological characteristics. ...
Soft computing strategies for resolving key data challenges in organ transplantation
(English) In the field of organ transplantation, a critical gap exists: the availability of organs falls far short of the demand, resulting in numerous recipients dying before they can receive a transplant. The complexity ...
Value engineering for autonomous agents
(English) The topic of this thesis is the engineering of values for autonomous agents. This is realised through the formulation, design and implementation of new functionalities for autonomous agents that enable reasoning ...
Towards a linearly organized embedding space of biological networks
(English) The recent technological advances in high-throughput sequencing have yielded vast amounts of large-scale biological omics data that describe different aspects of cellular functioning. These omics data are typically ...
DA&AI supporting tools for gas turbine’s efficiency improvement: maintenance, operation modes and performance enhancement
(English) Digitalization has revolutionized many industries, including the power generation sector. The availability of a vast amount of data from various systems has transformed decision-making processes in Industry. ...
Assessing biases through mosaic attributions
(English) Machine learning and, more specifically, deep learning applications have grown in number in recent years. These intelligent systems have shown remarkable performance across various domains, including sensitive ...
A methodology for the automation of building intelligent process control systems
(English) One of the major problems to design and implement control and supervision systems for real-world processes lies in the need to stablish an ad-hoc solution for each process, i.e., environmental systems in this ...
Causal discovery and prediction: methods and algorithms
(English) This thesis focuses on the discovery of causal relations and on the prediction of causal effects. Regarding causal discovery, this thesis introduces a novel and generic method to learn causal graphs by performing ...
New data integration methods for drug re-purposing by mining heterogeneous omics data
(English) High-throughput omics technologies produce large-scale heterogeneous data that provide complementary views of the underlying complexity of the studied organism. To exploit these new data to answer biomedical ...
A novel soft computing approach based on FIR to model and predict energy dynamic systems
We are facing a global climate crisis that is demanding a change in the status quo of how we produce, distribute and consume energy. In the last decades, this is being redefined through Smart Grids(SG), an intelligent ...
A novel computer Scrabble engine based on probability that performs at championship leve
The thesis starts by giving an introduction to the game of Scrabble, then mentions state-of-the-art computer Scrabble programs and presents some characteristics of our developed Scrabble engine Heuri. Some brief notions ...
A framework for the analytical and visual interpretation of complex spatiotemporal dynamics in soccer
Sports analytics is an emerging field focused on the application of advanced data analysis for assessing the performance of professional athletes and teams. In soccer, the integration of data analysis is in its initial ...
Language: universals, principles and origins
Here, old and new linguistic universals, i.e. properties obeyed by all languages on Earth are investigated. Basic principles of language predicting linguistic universals are also investigated. More precisely, two principles ...
Enhancing scene text recognition with visual context information
This thesis addresses the problem of improving text spotting systems, which aim to detect and recognize text in unrestricted images (e.g. a street sign, an advertisement, a bus destination, etc.). The goal is to improve ...
Deep learning architectures applied to wind time series multi-step forecasting
Forecasting is a critical task for the integration of wind-generated energy into electricity grids. Numerical weather models applied to wind prediction, work with grid sizes too large to reproduce all the local features ...
