Recent Submissions

  • Echoing the past, predicting the future: an exploration of various aspects related to Echo State Networks 

    Weber, Franziska (Universitat Politècnica de Catalunya, 2025-01-29)
    Master thesis
    Open Access
    Echo State Netzwerke (ESNs) sind eine Art rekurrenter neuronaler Netze bestehend aus einem großen, zufällig verbundenen rekurrenten Reservoir und einem anschließenden linearen Readout Layer. Nur die Gewichte des Readout ...
  • Extensions of infinite width neural network Kernels 

    Bosch Tobella, Guillem (Universitat Politècnica de Catalunya, 2024-10-24)
    Master thesis
    Open Access
    This thesis explores the diverse amount of known closed form expressions of the infinite width neural network kernels. These are kernels whose map to the feature space mimics the function defined by a single hidden layer ...
  • Modeling and predicting quality of service for application migrations in the cloud-edge continuum 

    Oliveras Torra, Joan (Universitat Politècnica de Catalunya, 2024-10-23)
    Master thesis
    Open Access
    In cloud-edge infrastructures, computing resources are distributed across centralized cloud servers and decentralized edge devices. This hybrid model allows for efficient resource utilization but also introduces challenges, ...
  • Deep learning for network anomaly detection in multivariate time-series 

    Xu, Ange (Universitat Politècnica de Catalunya, 2024-10-22)
    Master thesis
    Open Access
    This study evaluates state-of-the-art multivariate time series anomaly detection methods on Network Intrusion Detection System (NIDS) datasets converted into time series data. We find these methods perform well on attacks ...
  • Secured and conditioned generation of synthetic mammograms 

    López Rubio, Omar (Universitat Politècnica de Catalunya, 2024-10-23)
    Master thesis
    Open Access
    The use of generative models in medical imaging offers significant potential to address the scarcity of labeled medical data, particularly in mammography. The challenge we address in this work is to ensure that synthetic ...
  • Unsupervised categorization of killer whale pulsed calls 

    Paraskeva, Maria (Universitat Politècnica de Catalunya, 2024-10-22)
    Master thesis
    Open Access
    The identification and classification of marine mammal vocalizations play a crucial role in understanding their behavior and ecological interactions. In this thesis, we present an innovative approach to the unsupervised ...
  • Anomaly detection in firewall rules: a case study of CaixaBank's network 

    Puga Creus, Eduard (Universitat Politècnica de Catalunya, 2024-10-23)
    Master thesis
    Restricted access - confidentiality agreement
    This study examines anomaly detection within CaixaBank's firewall rule set, where complex interactions between rules can result in issues like redundancy, shadowing, and misconfigurations. Through the application of ...
  • Brute force vs. algorithmic approaches for chess board and pieces recognition 

    Medina I Diez, Adrià (Universitat Politècnica de Catalunya, 2024-10-23)
    Master thesis
    Restricted access - confidentiality agreement
    This thesis explores the integration of domain-specific knowledge into the design of machine learning (ML) systems, by examining the specific use case of chess board and pieces recognition. Using deep learning models, we ...
  • Automatic detection of data and concept drift in ML systems for chess broadcasting 

    Moure Alassio, Ximena Graciana (Universitat Politècnica de Catalunya, 2024-10-23)
    Master thesis
    Open Access
    As machine learning systems evolve, they face the challenge of operating in dynamic environments where model performance can degrade over time due to model drift. Therefore, the work continues once the model is in production; ...
  • A Graph Convolutional Network-based Framework for Federated Data Product Discovery 

    Bisquert Parés, Aniol (Universitat Politècnica de Catalunya, 2024-10-23)
    Master thesis
    Open Access
    Big data presents a novel opportunity for data-driven decision-making, and modern organizations acknowledge its potential. Being able to use these data collaboratively through advanced data-driven applications that use ...
  • Creating an AI-powered toolbox to analyse of the interaction betwen occupation and energy performance in commercial buildings 

    Méndez Serrano, Míriam (Universitat Politècnica de Catalunya, 2024-10-23)
    Master thesis
    Open Access
    This thesis aims to develop a robust digital architecture, dealing with the European meteorological data space called Destination Earth. Interactive and data-driven techniques will be used to analyse the characterisation ...
  • A Scalable, Multilingual, and Noise-Robust System for Music Metadata Extraction 

    Jácome González, Mateo (Universitat Politècnica de Catalunya, 2024-10-22)
    Master thesis
    Restricted access - confidentiality agreement
    The music industry is in an increasing need to develop tools and technologies to keep up with the development of the digital business landscape. The advent of user-generated content platforms and the record-breaking amount ...

View more