A Boxology of Design Patterns for Hybrid Learning and Reasoning Systems
This work provides a framework for researchers and practitioners in AI to clarify and reuse techniques in hybrid systems, but it is incremental as it builds on existing design pattern concepts from other fields.
The authors tackled the problem of systematizing hybrid learning and reasoning systems by proposing a set of compositional design patterns to describe combinations of statistical and symbolic techniques, and validated these patterns against recent literature.
We propose a set of compositional design patterns to describe a large variety of systems that combine statistical techniques from machine learning with symbolic techniques from knowledge representation. As in other areas of computer science (knowledge engineering, software engineering, ontology engineering, process mining and others), such design patterns help to systematize the literature, clarify which combinations of techniques serve which purposes, and encourage re-use of software components. We have validated our set of compositional design patterns against a large body of recent literature.