2023

Tikka, Santtu

Great strides have been taken in the field of causal inference with regards to identifiability of causal effects under various scenarios. Both experimental and observational data can be readily combined with state-of-the-art algorithms to answer causal questions that cannot be solved with any single data set alone. However, even though many of these algorithms are complete in a sense that they will always provide the correct solution and they fully characterize the problem at hand, there is still room for improvement from a practical standpoint. The thesis focuses on practical usability of causal identifiability algorithms. This summary provides a general overview of the thesis. For technical details and the discussed algorithms, the reader is referred to the main thesis and the cited papers. The thesis is available online at http://urn.fi/URN:ISBN:978-951-39-7519-7.