Abstract: 

In this session, we confront the widely acknowledged limitation in traditional statistical analysis and machine learning: 'correlation is not causation.' We start by dissecting this concept, outlining the challenges it presents when trying to derive meaningful insights from data.

The main focus then shifts to causal machine learning, a growing field that seeks to go beyond mere correlation and uncover actual cause-and-effect relationships in datasets. We discuss its importance when trying to make decisions on what to do based on predictions.

The talk further navigates through the appropriate conditions for applying causal machine learning, typically when randomized control trials are not possible due to ethical, logistical, or financial constraints. It outlines a range of sectors that could benefit from its application, including healthcare, economics, and social sciences.

To ground these concepts in reality, we present a practical use case, demonstrating the process of using causal machine learning to derive more actionable and robust insights.

Bio: 

Bernardo is a Data & AI leader, passionate about powering data transformation in companies and promoting social good in society using data.

He is specialized in Data Science, Machine Learning and AI, having won two awards in this field (Innovation in Big Data Award by Thomson Reuters and Machine Learning & Neural Computation Award by Imperial College London). His goal is to be able to take any challenge, no matter how complex, and to solve it using a fusion of art & science, business & technology capabilities, data & analytics to make it happen.

Bernardo has an MRes in Advanced Computing from Imperial College London, with a specialization in Machine Learning and a BSc in Electrical and Computer Engineering from Instituto Superior Técnico.