top of page

WHEN FEATHERS MEET ALGORITHMS: THE UNSEEN GENIUS OF PIGEONS

By P Mohan Chandran


Have you ever stopped to marvel at the navigational prowess of a homing pigeon or pondered the intricacies of bird intelligence? It seems that we have grossly underestimated our feathered cohabitants. The latest studies from Ohio State University and the University of Iowa reveal that the humble pigeon could hold the key to understanding Artificial Intelligence's most fundamental principles.


The ‘Brute Force’ of Avian Intellect


Picture this: a pigeon in a lab, faced with an array of visual cues – lines, rings, and patterns, embarking on a quest not unlike that of a sophisticated AI system. With each peck at a button, they make a choice, striving to sort these stimuli, rewarded by treats for their successes. Initially, their success rate mirrors a random guess, but with time, it soars to near perfection. This isn't a mere party trick. This is a window into cognition, a brute force method of problem-solving where trial and error reign supreme – a method strikingly similar to how AI learns and evolves.


Professor Brandon Turner, a psychology sage at Ohio State, finds a resounding parallel between pigeon learning mechanisms and the algorithms driving modern machine learning. It's a comparison that shakes the very foundation of what we perceive as 'intelligent' behavior.


A Lesson in Cognitive Flexibility


As the birds were subjected to these experiments, their ability to categorize, a cognitive feat we once believed was beyond their reach, shone through. They didn't just learn; they excelled, expanding their accuracy from a mere 55% to an astonishing 95% in more straightforward tasks. Such advanced cognitive and attentional processes showcased by these pigeons prompt us to reconsider the limits of non-human intelligence.


Artificial intelligence, crafted by human ingenuity, also navigates these tests, learning to curtail its errors. Yet, the AI’s approach, much like the pigeon’s, doesn’t stem from understanding but from building associations through persistent trial and error – a testament to the pigeons' efficient learning capability, albeit without the generalization skills that humans possess.


Turner illuminates a fascinating fact: associative learning, a method we associate with simpler minds, powers both pigeons and our most advanced technologies. This learning connects stimuli with outcomes, as seen in dogs who comprehend that sitting equals treats. We have long held the belief that such a primitive form of learning could not underpin complex tasks like visual categorization. The pigeons prove otherwise.


Beyond Rule-Based Thinking


Humans often abandon tasks when we cannot decipher the underlying rules. Not pigeons. They don’t seek rules; they thrive on the "brute force" of trial and error, surpassing humans in specific tasks. This method, though seemingly unsophisticated, may actually edge closer to the core of how learning and intelligence develop in nature and technology alike.


Turner reflects on the irony: humans exalt their intelligence for creating AI, yet often scoff at pigeons as simple-minded. The truth unveiled by this research is that the learning principles steering AI are not far removed from those utilized by pigeons.


As we stand at the crossroads of biological and artificial intelligence, one question looms large:


Could the key to the next AI breakthrough lie in the unassuming mind of a bird often disregarded by passersby?


And as the sun sets on the horizon, casting shadows on our preconceived notions of intelligence, we must ask ourselves:


What other secrets of cognition lie hidden in the natural world, waiting to be discovered not by surpassing nature, but by learning from it?

Do you agree that nature could uncover us many secrets which we are unaware?

  • A. Yes, nature has many mysteries unknown to humans.

  • B. No, we have already uncovered most secrets of nature.

© 2023. P Mohan Chandran. All Rights Reserved.

05112023



Comentarios


bottom of page