Below are videos demonstrating additional capabilities of Matheta AGI. The language used is from animal learning, which is what I was initially trying to model.
Instrumental Learning (Operant Conditioning)
A behavior followed by a reward increases the probability of that behavior. If followed by a punishment, it becomes less likely.
Classical (Pavlovian) Conditioning
An originally meaningless stimulus, when paired with a reflex, comes to elicit that reflex.
Reward and Punishment
How can a candy reward a machine? Put another way, how does an animal know it’s eaten? And how does getting food lead to remembering how it got the food?
A neutral stimulus is paired with a reward and thus becomes itself rewarding. Example: saying “Good Doggy” when giving a dog a snack.