TransferLab Seminar: Scientific Inference With Interpretable Machine Learning - Timo Freiesleben

The TransferLab Seminar (transferlab.ai) is a platform where researchers and engineers share and discuss recent advances in AI/ML, striking a balance between accessibility and mathematical depth.
Recorded talk from May 23rd, 2024.
Scientific Inference With Interpretable Machine Learning
Timo Freiesleben, postdoc at the Machine Learning in Science Cluster at the University of Tübingen
Abstract
To learn about real world phenomena, scientists have traditionally used models with clearly interpretable elements. However, modern machine learning (ML) models, while powerful predictors, lack this direct elementwise interpretability (e.g. neural network weights). Interpretable machine learning (IML) offers a solution by analyzing models holistically to derive interpretations. Yet, current IML research is focused on auditing ML models rather than leveraging them for scientific inference. Our work bridges this gap, presenting a framework for designing IML methods -termed ‘property descriptors’- that illuminate not just the model, but also the phenomenon it represents. We demonstrate that property descriptors, grounded in statistical learning theory, can effectively reveal relevant properties of the joint probability distribution of the observational data. We identify existing IML methods suited for scientific inference and provide a guide for developing new descriptors with quantified epistemic uncertainty. Our framework empowers scientists to harness ML models for inference, and provides directions for future IML research to support scientific understanding.

[@kevinhuhnstock836]

Voll cool 👍🏻👍🏻👍🏻