Short Trainings & Workshops

In this seminar we will introduce the basic concepts and building blocks behind Knowledge Graphs and Graph Neural Networks (GNNs). We will describe how they are used to make prediction about graph and entity properties and about relationship between entities. We will dive into some use cases where GNNs are used in the health domain, providing best practices and lessons learned from past projects.

This workshop provides an overview of techniques to model and predict propensity, namely the probability that a person will perform a given action in the future, based on past observations. It includes time-window classification and survival analysis, as well as customer lifetime value and causal analysis.

This workshop provides a general introduction to computer vision. It covers typical tasks where computer vision algorithms are employed, elaborationg on the basic principles and how to implement solutions (libraries, data prep, best practices,…). Such tasks include supervised and non-supervised learning (image recognition, object detection), transfer learning and diffusion models. At the end of the course we present a concrete use case from a company that uses computer vision to map & track marine plastic.

This workshop provides a high level overview of MLOps, focusing on the deployment of proof-of-concept AI solutions. We will introduce basic MLOps concepts and best practices, and present business use cases and corresponding specific recommendations.

This course offers an introduction to the recent developments of the Large Language Models. After a high level introduction about transformers, the course focus on LLMs and its applications. A special attention is given to ChatGPT and to the RLHF technique. The second part of the course is dedicated to a hands-on session (in Renku or Google Collaboratory) with examples and exercises on how to use models and libraries in Huggingface.

In this seminar the participants will learn what time series data is and its real-world applications. We will review time series key concepts, decomposition and the most popular forecasting methods. We will also explore clustering methods specifically designed for time series and we will conclude with some real world applications of time series analysis.

Explainability of AI models has become an important tool to deploy effectively many AI solutions, and to increase trust from the stakeholders of these solutions. In this workshop, we will introduce the most popular methods for explaining AI predictions, with a few examples of real use cases. Moreover, we will have a hands-on session with an exercise to try these techniques on a sandbox dataset.

This workshop will introduce participants to survival analysis, a technique originally used in medicine and now employed for analyzing the expected duration of time until any kind of event happens. During the workshop, participants will be introduced to the basic concepts behind survival analysis. They will then learn how to use them hands-on, by applying different algorithms on data relevant to practical use case, then evaluating them and extracting meaningful insights using Python.