This paper proposes adding probabilistic inference to a neuro-symbolic AGI robot framework based on Belnap’s typed intensional first-order logic, or IFOL_B. The stated goal is to assign probabilities to currently unknown sentences using Nilsson’s probability structure. It introduces a global symmetry transformation that preserves the current knowledge database and logical deductions, plus a local transformation for real-time decisions over a restricted subset of predicates. The probability density function KI is computed with neural networks using Shannon maximum-information-entropy principles. The abstract does not provide benchmark results or implementation details.
No heat snapshots are available in the last 24 hours.