Go to the Indico Home Page Oldest event Older event Up to category Newer event Newest event Printable version
 
Focus on:  Hide Contributions
  • Indico style
  • Indico style - inline minutes
  • Indico style - numbered
  • Indico style - numbered + minutes
  • Indico Weeks View
    • Switch to the management area of this event Back to Conference View
    Login

    Hands on Machine Learning

    Data controller
    IJS Data Controller
    dca@ijs.si
    Privacy notice
    → Europe/Ljubljana
    Amphitheater (Ajdovščina)

    Amphitheater

    Ajdovščina

    (Hybrid event - join in person or online)
    Saptashwa Bhattacharyya (UNG (Center for Cosmology and Astrophysics))
    Description

    Izvajalec / Course provider: University of Nova Gorica (UNG)
    Predavatelji / Instructors: G. Zaharijaš and S. Bhattacharyya

    Short course description

    This course is a hands-on microcredential focused on building robust ML analysis pipelines for real-world data. Participants progress from data preparation, regression and classification through model evaluation and cross-validation to deep neural networks and modern generative models. The course combines lectures with practical sessions. The course credits will be awarded to candidates who fulfill the attendance requirement (in person/online) and submit a short report/presentation detailing their methodology and practical application.

    Objectives and competencies

    • Build a robust data-analysis pipeline using machine-learning methods.
    • Apply the techniques to a broad range of real-world data, including tabular and unstructured data such as images.
    • Develop practical experience through hands-on exercises and a final project.

    Intended learning outcomes

    • Handle and process data before applying machine-learning methods.
    • Apply machine-learning algorithms to practical datasets.
    • Select, evaluate and compare ML models using appropriate performance metrics and validation procedures.
    • Apply the techniques learned in the course to a participant-selected use case.

    Course schedule

    Ten sessions, each 14:00–17:00:

    04/11/2026 Introduction to regression and classification
    06/11/2026 Performance metrics and evaluation
    11/11/2026 Building an analysis pipeline and cross-validation
    13/11/2026 Introduction to DNNs: backpropagation and DenseNet
    18/11/2026 Analysis with DenseNet: tabular and image data
    20/11/2026 Introduction to CNNs and different architectures
    25/11/2026 Image classification and segmentation with CNNs
    27/11/2026 Variational Autoencoder and GAN
    02/12/2026 Probabilistic diffusion models
    04/12/2026 Generation and evaluation of images

     

    Learning and teaching methods

    • Lectures combined with hands-on sessions.
    • Exercises applying different methods and evaluating the performance of ML models.
    • Final project: construction of a complete ML analysis pipeline for a particular task using a dataset chosen by the participant.

    Assessment

    • Class attendance: 35%
    • Final project and presentation: 65%

    Readings

    • Deep Learning (Y. Bengio et al.) – available online.
    • Stanford machine-learning course notes – available online.
    • Understanding Deep Learning (S. Prince) – available online.
    Contact
    Registration
    General registration form for SLAIF
    Register