In order to alleviate the task of logical consistency checking in Machine Leaning, and thus facilitate the training process of Language Models, we are developping solution for contrained ML inference and training to a defined formal setting.

Merging neuromorphic computing with interaction combinators, we develop solution to simulate brain-like processes at extremely low overhead and high flexibility. This work is partly supervised by our Ethics Lab, as part of our commitment to moral maximization. We plan to launch the happy fly™ project by 2026, providing a platform for simulation up to 10¹⁴ fruit flies, enabling unprecedented good manufacturing.