This tutorial introduces federated learning, including its core algorithm (FedAvg), privacy and security challenges (e.g. Byzantine workers), and recent advances in personalization and efficiency. An overview of economic motivations and practical deployment considerations is also included.

The tutorial covers strategies for learning with limited labeled data in medical imaging, including synthetic data generation and domain adaptation. It also introduces implicit neural representations as a tool for flexible image reconstruction.

This session examines the extent to which large language models (LLMs) learn and represent linguistic knowledge. It considers experimental evidence from training under extreme conditions and reflects on the implications for language understanding and task performance.

This session outlines key concepts in quantum computing and explores how machine learning methods are used within quantum workflows. It also introduces the field of quantum-assisted machine learning, focusing on current capabilities and open questions.

A review of developments in vision-language models and self-supervised learning. The tutorial surveys architectures such as CLIP and Segment Anything, with attention to pretraining methods, embedding alignment, and open-vocabulary tasks.

This session presents applications of AI in analyzing genomic data for cancer research. Methods covered include traditional ML, deep learning, and large language models for detecting mutations and identifying therapeutic targets.