Growing Personality inside a machine.
Posted: Fri Dec 18, 2020 5:50 amPsychometrics have been around for decades and inform us about the limits to our intelligence and the traits desirable or not of our many varied personality's. But what if we could delve deeper and allow our measurements to go beyond mere types. Researchers at the Stanford institute have done just that inside a machine simulation.
Two AIs exist each having been primed by the measurements taken from two people over the course of several years and some dialogue taken from Albert and Costello. Inside a simu!action they simply play out a meaningful dialogue over many billions of iterations. Each word in the dialogue is pretrained to its base meaning. The result is two distinct auto generative artificial personalitys.
The absence of big data training has meant that there is no margin for error as every word in meant and has a context to which the AI's are only too well aware.
When isolated and combined within its hierarchical system the emergence of friendly and helpful robots combined with task completion becomes a winning strategies that powers the mars habitat building robotic fleet of tommorow.
Two AIs exist each having been primed by the measurements taken from two people over the course of several years and some dialogue taken from Albert and Costello. Inside a simu!action they simply play out a meaningful dialogue over many billions of iterations. Each word in the dialogue is pretrained to its base meaning. The result is two distinct auto generative artificial personalitys.
The absence of big data training has meant that there is no margin for error as every word in meant and has a context to which the AI's are only too well aware.
When isolated and combined within its hierarchical system the emergence of friendly and helpful robots combined with task completion becomes a winning strategies that powers the mars habitat building robotic fleet of tommorow.