Affective Neuron™ ~ Novel AI
Affective Neuron™ is a new patent-pending unsupervised reinforcement learning synthetic neuron.
In meeting its objective - to minimize the need to expend energy - a side effect involves the minimization of stress. Don't we aim for that?! Affective Neuron™ is conditioned not trained; is unbiased and learns in real-time, continuously.
Unlike all current neural network machine learning technologies, Affective Neuron™ is not based upon the connectionist model of cognitive processing. The central connectionist principle is that mental phenomena can be described by interconnected networks of simple and often uniform units.
The Affective Neuron™ takes the opposite approach, where neurons are heterogenous complex processing elements capable of storing and sharing complete memories. The connectionist model is over 30 years old.
Recent experiments on brain cells demonstrate that neurons are complex processing elements. This is opposite to the connectionist model of cognition. Engrams are a hypothetical permanent change in the brain accounting for the existence of memory (engrams are memory traces). Affective Neurons can automatically import human engrams to reflect these memories in their processing. Affective Neurons are interconnected through novel dynamic quantum diffusion processing networks, which mimics how biological neuron synaptic interfaces function.
This allows Affective Neurons to migrate and evolve complex thought processing networks, and to share their diverse memories to support sophisticated thoughts and understandings.