26 May 2025 to 6 June 2025
UNIST Technological Park, Croatia
Europe/Zagreb timezone

Scientific Programme

The Split AI Lab is an intensive two-week program designed to provide participants with a strong foundation in Machine Learning and Data Science, combining theoretical instruction with hands-on implementation. The curriculum is structured to gradually build knowledge, covering fundamental statistical concepts, essential machine learning algorithms, and modern deep learning techniques.

The program is divided into three main components: supervised learning, unsupervised learning, and deep learning, with each section emphasizing both theoretical understanding and practical application.

Core Topics Covered
Statistical Foundations & Data Preprocessing
essential statistical concepts and learning best practices in feature engineering, dimensionality reduction, and data preprocessing techniques.

Supervised Learning
classification and regression models, including k-Nearest Neighbors (kNN), Decision Trees, Linear and Logistic Regression, Support Vector Machines (SVMs), and Ensemble Methods. Emphasis is placed on model evaluation, cross-validation, and trade-offs between different performance metrics.

Unsupervised Learning & Clustering Techniques
k-Means Clustering, Principal Component Analysis (PCA), Linear Discriminant Analysis (LDA), and Singular Value Decomposition (SVD), along with methods for analyzing high-dimensional data.

Advanced Topics & Model Optimization
kernel methods, regularization techniques, non-linear regression models, and an introduction to structured machine learning models.

Deep Learning & Neural Networks
fundamental neural network architectures, including feedforward networks, convolutional neural networks (CNNs), and basic deep learning frameworks.

Hands-on Learning Approach
Throughout the program, participants will implement these models using Jupyter Notebooks and standard Python libraries such as NumPy, Pandas, Scikit-learn, and TensorFlow/PyTorch. Practical applications will include real-world datasets and Kaggle-style challenges, allowing participants to gain experience in model training, tuning, and performance evaluation.