The answer to a specific problem generally involves coupling a series of such algorithms. Depending on requirements this coupling can be made directly at Machine Learning level, or via a higher-level structuring layer, for example a workflow applying black boxes of individual learning algorithms in stages. Machine Learning algorithms can also be considered as low-level algorithms allowing classification on concepts of a higher-level structure enabling formal semantics: logic description reasoner, theorem demonstrator, conceptual graph, etc. Once enhanced in the semantic layer, the integrated inference engine will be able to couple the concepts obtained to make inferences similar to human reasoning.
In the first instance learning is done with a ‘master’ black box formed by coupling the algorithms comprising the original black boxes. The interest of this combination in a high-level black box is, for example, to obtain - after training - classification capabilities superior to those of an isolated black box, generally by identification of a chain of command of reasons that each layer processes at its own level. However we lose the capacity to control and communicate the ‘good practices’ identified by these closed systems. As a result this system is very close to human intuition, which makes it possible to accelerate the choice processes in complex environments.
In the second case, compatible with the strategy of the first but located at a higher level, the overall structuring of the system will be based on a workflow engine coupled with a structure capable of self-organizing knowledge at a level close to formal human structuring. This structuring is a typical prerequisite for exchanging practices and intuitions with counterparts and constructing advanced organizations.
This two-tier application optimization strategy is the cornerstone of the Big/Smart Data methodology and toolkit provided by logiCells application server and framework. The latter helps tackle complex business issues, and gives them the flexibility required to survive in a world where the pace of change is constantly accelerating.