Synthetic Philosophy and Deductive Engineering
The role of the deductive intelligence subsystem is to enhance the capabilities of the deductive knowledge subsystem by implementing advanced reasoning strategies. It leverages techniques such as alpha-zero style reinforcement learning optimized Monte Carlo Tree Search (MCTS) to guide the proof search process efficiently.
The core thesis underlying this capability is that every logical context is a perfect information system and is therefore amenable to such techniques, and that the heirarchical nature of the logical contexts (in which each context is obtained by exending a smaller context ultimately rooting in the core context of the pimitive logical system) lends itself to the exploitation of heuristics implenented through heirachically layered neural nets.
In effect this subsystem implements specialists for each logical context, but the architecture of SPaDE allows for third parties to compete in providing specialist capabilities both in terms of the knowledge applied and in terms of the means of exploitation. It is intended that this will encourage the development and application of verified design algorithms which yield proven solutions to engineering (or other) problems.