Modular design patterns for neural-symbolic integration: refinement and combination
This work addresses the need for structured integration of neural and symbolic AI, but it is incremental as it builds on existing patterns.
The paper formalizes neural-symbolic design patterns to define refinement and modular combination, implementing these in the heterogeneous tool set (Hets) for checking well-formedness and computing combinations.
We formalise some aspects of the neural-symbol design patterns of van Bekkum et al., such that we can formally define notions of refinement of patterns, as well as modular combination of larger patterns from smaller building blocks. These formal notions are being implemented in the heterogeneous tool set (Hets), such that patterns and refinements can be checked for well-formedness, and combinations can be computed.