Knowledge representation and discovery with strong-connected networks

We propose strong-connected networks (SCN) for knowledge representation and discovery. Knowledge is represented among nodes and connections of SCN. Knowledge discovery in SCN emulates natural brain-like processes with dynamic label-free connections and the additional external unit. This method is used in a tabular data domain and shows advantages in table order insensitivity, flexibility, and speed. Moreover, stochastic property in this process adds uncertainty in knowledge discovery, which might be similar to cognitive knowledge discovery process.