Large funding from BMBF to develop causal ML

Enlarge

The "CausalNet" project aims to develop novel ways to integrate causality into ML models with partners from KIT, Helmholtz Munich, and Economic AI.

Existing machine learning models typically rely on correlation, but not causation. This can lead to errors, bias, and eventually suboptimal performance. In our project "CausalNet", we aim to develop novel ways to integrate causality into ML models. Concretely, our focus is on:

• Flexibility: We develop a general-purpose causal ML model that can handle a broad range of settings, including high-dimensional, time-series, and multi-modal data.
• Efficiency: We develop techniques for efficient learning algorithms (e.g., synthetic pre-training, transfer learning, and few-shot learning) that are carefully tailored to causal ML.
• Robustness: We create new environments/datasets for benchmarking. We also develop new techniques for verifying and improving the robustness of causal ML.Open-source: We fill white spots in the causal ML toolchain to improve industry uptake.Real-world applications: We will demonstrate performance gains through causal ML in business, public policy, and bioinformatics for scientific discovery.

Other participating members are Prof. Nadja Klein (KIT), Prof. Stefan Bauer (TUM / Helmholtz Munich), Prof. Niki Kilbertus (TUM / Helmholtz Munich), and Economic AI GmbH.