How Strong AI recognizes an image
The Ghost Perl AI needs some brave soul (and Perl coder :-) to implement the VisRecog mind-module depicted in the above animation, which I created on the Commodore Amiga 1000 computer in my relative youth. (Thank you to whoever republished it as a public “gif” image.)
Above we see an AI Mind thinking in Latin (something like “Cave Canem” or “Beware of Dog”) while seeing and recognizing a dog. An old image of a dog tickles or activates an associative tag over to the concept of “canis” in Latin or “dog” in English. Then the concept sends a spike across an associative tag over to the Latin word-stem “can-” for “dog”. The AI software then adds an inflectional ending to form the accusative singular word “Canem”.
Back in the visual memory channel, the newly recognized image of a dog gets recorded as one more instance of the concept of “dog” — tagged above as the most recent “percept” of a dog. The associative pathways are bi-directional. We can hear “Cave canem” or “Beware of dog” and conjure up an image of a ferocious animal like Cerberus guarding the underworld. Or we can see a dog and think an idea (“light-bulb” above) in response to the perception.
Since I created the animation on my Amiga and therefore I own the rights to the image, I will publish this article in the public domain and I hereby grant Netizens total freedom to use my VisRecog animation in their AI work rushing towards the impending Technological Singularity. I also urge AI enthusiasts to download the free AI source code and Strawberry Perl 5 so as to install a personal AI Mind on their own computer and on all computers of other Singularitarians wishing to participate in the evolution of artificial intelligence.