I wonder if extremely well trained dogs might work?
Chaser seems likely to have learned nouns, names, verbs… with toy names learned on one trial starting at roughly 5 months of age (albeit with a name forgetting curve so additional later exposures were needed for retention).
Having studied her training process, it seems like they taught her the concept of nouns very thoroughly.
Showing “here are N frisbees, after ‘take frisbee’ each one of them earns a reward” to get the idea of nouns referring to more than one thing demonstrated very thoroughly.
Then maybe “half frisbees, half balls” so that it was clear that “some things are non-frisbees and get no reward”.
I think that sort of task might be modellable with ACT-R—the hardest part might be getting or gathering the animal data to compare to! Most of the time ACT-R models are validated by comparing to human data gathered by taking a room full of undergraduates and making them do some task 100 times each. It’s a bit trickier to do that with animals. But that does seem like something that would be interesting research for someone to do!
I wonder if extremely well trained dogs might work?
Chaser seems likely to have learned nouns, names, verbs… with toy names learned on one trial starting at roughly 5 months of age (albeit with a name forgetting curve so additional later exposures were needed for retention).
Having studied her training process, it seems like they taught her the concept of nouns very thoroughly.
Showing “here are N frisbees, after ‘take frisbee’ each one of them earns a reward” to get the idea of nouns referring to more than one thing demonstrated very thoroughly.
Then maybe “half frisbees, half balls” so that it was clear that “some things are non-frisbees and get no reward”.
In demos of names and verbs, after the training, you can watch her looking at things and thinking. Maybe the looking directions and the thinking times could be modeled?
I think that sort of task might be modellable with ACT-R—the hardest part might be getting or gathering the animal data to compare to! Most of the time ACT-R models are validated by comparing to human data gathered by taking a room full of undergraduates and making them do some task 100 times each. It’s a bit trickier to do that with animals. But that does seem like something that would be interesting research for someone to do!