The Science of QAI

Below (and above) is a video 3D screen capture of Quantum Artificial Intelligence (QAI) signaling neuron superposition viewed in the QAI cognitive reactor. QAI signaling agents route messages enabling network node collective intelligence. Briefly, there are only two quantum signaling agents shown in the video. One of the two learns using new patent pending quantum reinforcement learning technology, and the other just responds without learning. The learning signaling neuron is demonstrated to replicate the results of Young's two slit experiment being in two places simultaneously. This is not two copies of the signaling neuron nor a copy shared, but a single signaling neuron that is in a state of coherent superposition simultaneously functioning in two (or more) places at the same time. The value of this is that QAI emergent deep learning signaling networks are composed of billions of signaling agents in quantum superposition, i.e. a Schrödinger's Cat.

QAI machine learning implements symmetry breaking to exit out of local minimum solutions to achieve the global optimum within its emergent deep learning signaling networks.

QAI Neuron Coherent Superposition

In physics, symmetry breaking is a phenomenon in which (infinitesimally) small fluctuations acting on a system crossing a critical point decide the system's fate, by determining which branch of a bifurcation is taken. To an outside observer unaware of the fluctuations (or "noise"), the choice will appear arbitrary, but it's not.

Symmetry breaking transforms a particle system from one where there are many local optimum false solutions to a system with more definite states and the global optimum. Our current universe is an example. The human brain is another example where symmetry breaking occurs to rapidly achieve learning objective stasis.

NP-Complete problems are exponential and only approximate solutions are possible, where a specific solution to a well bounded NP-Complete problem is achieved. An example is the "chip" layout problem for integrated circuits. Machine learning problems are NP-Complete. This has been documented by Sutton et al. Ergo, only narrow solutions are possible with all current AI and machine learning.