In a fascinating intersection of natural and artificial intelligence, a recent study has demonstrated that pigeons, often dismissed as “dim-witted,” actually employ learning strategies like those used in AI to solve complex tasks.
This new research highlights the capabilities of associative learning in overcoming challenges that would typically leave humans frustrated.
Researchers at The Ohio State University led by Professor Brandon Turner, along with Professor Edward Wasserman from the University of Iowa, have found evidence that pigeons can navigate difficult categorization tasks that defy traditional human cognitive strategies like selective attention and explicit rule application.
The researchers examined the problem-solving methods used by pigeons and tested a simple AI model to see if it could solve the problems in the same way, and it worked.
“We found really strong evidence that the mechanisms guiding pigeon learning are remarkably similar to the same principles that guide modern machine learning and AI techniques,” said Professor Turner.
“Our findings suggest that in the pigeon, nature may have found a way to make an incredibly efficient learner that has no ability to generalize or extrapolate like humans would.”
The research involved a series of experiments where pigeons were trained to categorize stimuli, including various lines and rings, by pecking a button corresponding to the correct category. Successful choices were rewarded with food pellets, encouraging the birds to improve their accuracy through trial and error.
Notably, the pigeons showed significant improvement in their ability to categorize the stimuli, even in more demanding tasks, indicating that their learning process was not as primitive as previously assumed.
The birds appeared to utilize associative learning, a fundamental form of making connections between different elements, much like dogs being trained to associate a command with a treat.
This form of learning was previously thought to be too rudimentary to account for the complex visual categorization displayed by the pigeons.
However, the AI model, based on associative learning and error correction – presumed to mirror the pigeons’ learning mechanisms – was also able to tackle the tasks effectively, further reinforcing the researchers’ proposition.
The study results challenge the common perception of human superiority in creating complex AI systems by paralleling the learning principles of artificial intelligence with those observed in pigeons.
Facing the same challenges that were presented to pigeons, humans would try to come up with a rule or rules that could make the task easier, noted Professor Turner.
“But in this case, there were no rules that could help make this any easier. That really frustrates humans and they often give up on tasks like this,” he said.
“Pigeons don’t try to make rules. They just use this brute force way of trial and error and associative learning and in some specific types of tasks that helps them perform better than humans.”
What’s interesting, though, is that pigeons use this method of learning that is very similar to AI designed by humans, said Turner.
“We celebrate how smart we are that we designed artificial intelligence, at the same time we disparage pigeons as dim-witted animals. But the learning principles that guide the behaviors of these AI machines are pretty similar to what pigeons use.”
The research, supported by from the National Science Foundation and the National Institutes of Health, raises compelling questions about the underestimated cognitive abilities of animals.
The study is published in the journal iScience.
Like what you read? Subscribe to our newsletter for engaging articles, exclusive content, and the latest updates.