Smart Machine Learning

In this work, a cognitive computing model, a cognitive computing architecture, and a cognitive computing system are proposed and tested to address a big data classification problem. The proposed cognitive computing model is called the STE-M model and it adopts the standard components, senses (S), thoughts (T), experiences (E), and memory (M), of human cognition to describe the processes involved in cognitive computing for big data classification. Similarly, the proposed cognitive computing architecture is called the cognitive random forest and it amalgamates the STE-M model and a set of random forest classifiers to enhance continuous learning. It also includes intra- and intercognitive computing models to connect STE-M and random forest models and improve classification accuracy with spatial and temporal reasoning. A cognitive computing system is also proposed and it is used to validate the proposed cognitive computing architecture. Experiments with a robotic navigation scenario under different environmental conditions show that the proposed cognitive random forest is capable of handling the environmental conditions what we called the cune conditions for big data.