Mandombe-Based Artificial Intelligence: Toward a Symbolic-Geometric Model of Computation

Résumé

This paper focuses on adapting Mandombe's recursive symmetry to artificial intelligence design. It proposes a symbolic-geometric learning algorithm that uses inversion and rotation logic instead of linear weighting. Simulations demonstrate higher interpretability and resistance to data bias, aligning with decolonial computing principles.