Cyber Valley Research Fund improves human decision-making with AI
Scalable machine learning for future challenges
A research project funded by the Cyber Valley Research Fund focused on using machine learning to improve human decision-making. This innovation was made possible by unrestricted resources and support for scientific research provided by six of Cyber Valley’s founding partners: Amazon, BMW, Bosch, IAV, Porsche, and ZF.
From February 2020 to February 2024, Dr. Falk Lieder led a major project supported by the Research Fund, titled “A Scalable Machine Learning Approach to Improve Human Decision Making,”. The project aimed to prepare for a future where effective decision-making is crucial.
Good decision-making involves planning, gathering relevant information, and handling uncertainty. As automation transforms the workforce by taking over routine tasks, the demand for jobs requiring decision-making in complex and novel situations will rise. Future roles will emphasize critical thinking and adaptability. To prepare, society must focus on retraining workers and equipping the next generation with advanced decision-making skills.
Lieder’s project leveraged artificial intelligence to uncover optimal decision strategies that help individuals navigate complex challenges. The team also designed and tested interactive tutorials, allowing participants to learn and apply these strategies effectively.
They evaluated their approach on a series of more realistic decision problems, ranging from an abstract planning task without uncertainty, to planning under uncertainty, to a company choosing between alternative projects. In each case, the new AI methods found decision strategies that outperformed both human decision-making and strategies identified by previous techniques.
Their interactive tutorials improved decision-making by explaining the strategy in clear language, demonstrating it through examples, and offering hands-on practice with feedback. In all three applications, individuals who completed the AI-driven strategy tutorials made better decisions than those who practiced decision-making independently.
This project yielded the following peer-reviewed publications:
-
Consul, S., Heindrich, L., Stojchewski, J., & Lieder, F. (2022), Improving Human Decision-Making by Discovering Efficient Strategies for Hierarchical Planning, Computational Brain & Behavior, 5(2), 185-216. https://doi.org/10.1007/s42113-022-00128-3
-
Callaway, F., Jain, Y. R., van Opheusden, B., Das, P., Iwama, G., Gul, S., Krueger, P.M., Becker, F., Griffiths, T. L., & Lieder, F. (2022). Leveraging Artificial Intelligence to Improve People's Planning Strategies. PNAS, 119 (12) e2117432119. https://doi.org/10.1073/pnas.2117432119
-
Mehta, A., Jain, Y. R., Kemtur, A., Stojcheski, J., Consul, S., Tosic, M., & Lieder, F. (2022). Leveraging machine learning to automatically derive robust decision strategies from imperfect knowledge of the real world. Computational Brain & Behavior, 5(3), 343-377. https://doi.org/10.1007/s42113-022-00141-6
-
Becker, F., Skirzynski, J., van Opheusden, B., & Lieder, F. (2022). Boosting human farsightedness with automatically discovered planning instructions. Computational Brain & Behavior, 5(4), 467–490. https://doi.org/10.1007/s42113-022-00149-y
-
Skirzynski, J., Jain, Y. R., & Lieder, F. (2024). Automatic discovery and description of human planning strategies. Behavior Research Methods, 56(3), 1065-1103. https://doi.org/10.3758/s13428-023-02062-z
-
Skirzyński, J., Becker, F., Lieder, F. (2021). Automatic Discovery of Interpretable Planning Strategies. Machine Learning, 110, pp. 2641–2683. https://doi.org/10.1007/s10994-021-05963-2
-
Kemtur, A., Jain, Y. R., Mehta, A., Callaway, F., Consul, S., Stojcheski, J., & Lieder, F. (2020). Leveraging machine learning to automatically derive robust planning strategies from biased models of the environment. Proceedings of the 42nd Annual Meeting of the Cognitive Science Society.
Preprints of articles yet to be published:
-
Heindrich, L., Consul, C. & Lieder, F. (2024). Leveraging AI to improve human planning in large partially observable environments. arXiv Preprint. https://arxiv.org/abs/2302.02785
-
Heindrich, L., & Lieder, F. (2024). Leveraging automatic strategy discovery to teach people how to select better projects. https://arxiv.org/abs/2406.04082
Dr. Falk Lieder is now an Assistant Professor of Psychology at UCLA and the Director of the Rational Altruism Lab.