MaLGa logo MaLGa black extended MaLGa white extended UniGe ¦ MaLGa UniGe ¦ MaLGa Universita di Genova | MaLGa Universita di Genova

Machine Learning for Data Science

  • Our main focus is to apply machine learning to real-world data, making sure to design stable, reproducible and interpretable applications with an impact in real-life scenarios.

    To this aim, we design and use machine learning methods that are based on:

    • regularization methods
    • deep learning
    • statistical modeling
  • The main goal is to analyze complex, temporal, and possibly structured data in order to:

    • identify predictive signatures
    • discover interpretable patterns
    • infer and analyze networks

Research

  • Network analysis and inference for websites

    We exploit statistical modeling and artificial intelligence to redesign large-scale websites.

    Studying the structure of large websites

    We describe the empirical properties of large-scale websites with inferential statistics and random models.

    Topic modeling from structured texts of websites

    Given a large collection of structured texts, we infer the network based on topic similarities through topic modeling techniques combined with deep learning representations.

    Collaboration with: Andrea Vian (UniGe)

    • AI-driven web design
    • AI-driven web design
  • New directions

    We are investigating machine learning methods for time series analysis in the contexts of:

    Meteorology

    • wave height forecasting
    • wind and rainfall nowcasting from underwater measurements

    Collaboration with: Agnese Seminara (CNRS, France), Giovanni Besio (UniGe), Andrea Trucco (UniGe)

    Finance

    • Risk assessment of defaulting
    • Target price forecasting for energy markets