The Synchem systems is a large knowledge-based domain specific heuristic problem solving program that is able to find valid synthesis routes for organic molecules of substantial interest and complexity without online guidance on the part of its user. Synchem requires an extensive knowledge base to make it routinely useful. However, it is very difficult to engage domain experts to the long-term dedication and intensity of commitment necessary to create a genuinely productive knowledge base. In order to deal with the debilitating knowledge-base bottleneck, machine learning programs will be devised that can use large computer- readable databases of specific reaction instances to provide training examples for algorithms designed to extract the underlying reaction schemata through inductive and deductive generalization. This approach will be augmented by the methodology that is usually described as explanation-based learning. Since the individual reaction entries in most databases are often haphazardly sorted and classified, another machine learning program will partition the input databases into coherent reaction classes using the methodology of conceptual clustering. To make it possible for Synchem to deal with the explosion of reasonable pathways that would be the result of a successful effort to greatly expand the knowledge base, the inference engine will reformulated for distributed parallel search of the problem space.
