Jakob Zeitler is a Pioneer Fellow at SMARTbiomed and a PhD graduate from the UCL Centre for Artificial Intelligence. In this conversation, we explore the intersection of causal inference and computer science, focusing on its theoretical foundations and practical applications. We discuss how frameworks like Directed Acyclic Graphs (DAGs) and potential outcomes are used to model causality, the differences between observational and experimental studies, and the growing role of causal inference in AI and machine learning. The conversation includes real-world examples, industry use cases, and guidance for applying causal methods in both academic and business settings.
Expect to learn about
-
- The fundamentals of causal inference, including DAGs and potential outcomes.
- The practical differences between observational and experimental studies.
- How causal reasoning is applied in fields like public health, forecasting, and logistics.
- The challenges and opportunities of integrating causal inference with machine learning.
- Methods such as proximal learning and synthetic control.
- The limitations of causal discovery tools in business contexts.
- How industries are leveraging experimentation and causal inference for decision-making.
Where to find Jakob
Website: https://jakobzeitler.github.io/
LinkedIn: https://www.linkedin.com/in/jakobzeitler/
Timestamps
00:00 Preview
00:46 Intro
01:24 Causal inference in Computing
04:22 DAGs vs. potential outcomes
07:59 Proximal learning
16:34 Theory vs. practice
20:18 Making assumptions & observations
26:32 The cost of assumptions
33:26 Including disclaimers
36:16 Causality in forecasting
41:54 What applied problems do
43:33 Causal discovery
51:07 CS vs. economist training
59:25 Causal inference in everyday life