A method for a continuously learning and optimizing artificial intelligence (AI) adaptive neural network (ANN) computer comprising monolithic microwave integrated analog circuits (MMICs) and incorporating a Cerebellar Model Articulation Controller (CMAC) neural network for human affect computer modeling (HACM), for predicting, modeling, and simulating human reactions to repeated human stress and/or affect responses to sensory inputs from both external and internal environmental stimuli, the method using learning control and multiple parallel processing neural network pathways, and not using pattern recognition, the AI ANN modeling method comprising:
- (a) executing at least one sensory input collection module on at least one computer system using non-transitory computer readable media operative to collect sensory input data representing at least one set of external environment stimuli inputs and at least one set of internal environmental stimuli inputs from at least one internal environment and at least one external environment;
- (b) executing at least one recording module on at least one computer system using non-transitory computer readable media operative for recording the sensory input data, wherein the sensory input data corresponds to at least two sensory input-specific features;
- (c) executing at least one context module on al least one computer system using non-transitory computer readable media operative to identify and generate at least one current context AI ANN model that processes said sensory input data to provide human reaction model data that is configured for use in a human reaction modeling module that predicts, simulates and models said simulated human reactions to said repeated human stress and/or affect responses to said sensory inputs from said both external and internal environmental stimuli;
- (d) executing at least one vector generating module on at least one computer system using non-transitory computer readable media operative to process the human reaction model data to generate position locator and track data comprising velocity vectors and acceleration vectors that determine at least one position locator on at least one corresponding position locator track traversing at least one dimension, wherein said position locator and track data correspond to said human reaction model data;
- (e) executing at least one human reaction modeling module on at least one computer system using non-transitory computer readable media, the human reaction modeling module operative to use said position locator and track data that is continuously updated to generate human reaction model data that predicts, simulates and/or models said simulated human reactions to said repeated human stress and/or affect responses to said sensory inputs from said both external and internal environmental stimuli;
- (f) executing at least one cognitive learning module on at least one computer system comprising monolithic microwave integrated analog circuits (MMICs) using non-transitory computer readable media operative to:
- (i) compare said predictive reaction/behavior data, corresponding to said predicted, simulated and/or modeled simulated human reactions, with corresponding measured reaction/behavior data from measured or recorded stress and/or affect parameter data of comparative human subjects in similar internal and/or external environments and current contexts;
- and
- (ii) generate continuously updated predictive reaction/behavior data from said human reaction model data and from said measured reaction/behavior data using at least one MMIC,
- wherein said at least one cognitive learning module predicts, simulates and models said simulated human reactions to said repeated human stress and/or affect responses using said position locator and track data and said measured reaction/behavior data;
- and wherein;
- and
- (g) executing at least one optimizing module on at least one computer system using at least one MMIC;
- and using non-transitory computer readable media operative to generate optimized predictive reaction/behavior data using first and second derivative optimization of said position locator and track data with learning control and multiple parallel processing neural network pathways to determine optimized sets of said velocity vectors and acceleration vectors corresponding to optimized position locator and track data that provide said optimized predictive reaction/behavior data with the highest correspondence to said measured reaction/behavior data and/or data extrapolated therefrom.