Deliverable D5.7 explores how ethical and cultural values can be formally embedded into AML-based world models. It shows how user interactions and situational consequences help autonomous systems learn to behave in socially and culturally acceptable ways. Two short demo videos illustrate these methods in action, offering practical examples of ethical alignment in decision-making.
The ALMA project has released Deliverable D5.7, which explores how ethical and cultural values can be embedded into machine reasoning using Algebraic Machine Learning (AML).
This work demonstrates how AML models can adapt based on user decisions and cultural contexts, ensuring that autonomous systems align more closely with human expectations and norms. Two demonstration videos illustrate key contributions from the deliverable:
-
Simulating Ethical and Cultural Consequences (Section 4.4)
How user-defined ethics influence driver behavior and outcomes in a real-time simulation.
-
Retraining AML Models from User Interaction (Section 4.6)
How AML uses prior experience to refine future recommendations, such as the placement of safety cameras.
This work supports ALMA’s broader goals of trustworthy Human–AI co-learning and responsible autonomy.
🔗 Explore the open-source AML engine, documentation, and examples:
https://github.com/Algebraic-AI/Open-AML-Engine
MORE INFORMATION ABOUT ALMA:
For any questions please contact
This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 952091.