How can baby learning help teach AI?


Babies can help unlock the next generation of artificial intelligence (AI), according to neuroscientists and colleagues at Trinity College, who just published new guiding principles for improving AI.

The study released today [Wednesday 22 June 2022 ] in the diary Nature Machine Intelligenceexamines the neuroscience and psychology of child learning and distills three principles to guide the next generation of AI that will help overcome machine learning’s most pressing limitations.

dr Lorijn Zaadnoordijk, Marie Skłodowska-Curie Research Fellow at Trinity College explained: “Artificial intelligence (AI) has made tremendous strides in the past decade, giving us smart speakers, autopilots in cars, ever smarter apps, and improved medical diagnostics. These exciting developments in AI have been achieved thanks to machine learning, which uses huge data sets to train artificial neural network models. However, progress is stagnating in many areas because the data sets that machines learn from have to be painstakingly curated by humans. But we know that learning can be done much more efficiently because infants don’t learn that way! They learn by experiencing the world around them, sometimes even by seeing something just once.”

In her article “Lessons from Infant Learning for Unsupervised Machine Learning”, dr Lorijn Zaadnoordijk and Professor Rhodri Cusack from Trinity College Institute of Neuroscience and Dr. Tarek R. Besold from TU Eindhoven, Netherlands, argue that better methods of learning from unstructured data are needed. For the first time, they make concrete suggestions as to which special findings from the learning of small children can be fruitfully applied in machine learning and how these findings are to be applied exactly.

Machines, they say, need built-in preferences to shape their learning from the start. They must learn from richer datasets that capture how the world looks, sounds, smells, tastes and feels. And like infants, they must go through a developmental path in which experiences and networks change as they “grow up.”

dr Tarek R. Besold, Researcher, Philosophy & Ethics Group at TU Eindhoven, said: “As AI researchers, we often draw metaphorical parallels between our systems and the mental development of human babies and children. It is high time to take these analogies more seriously and look at the wealth of knowledge about child development from psychology and neuroscience that can help us overcome the most pressing limitations of machine learning.”

Professor Rhodri Cusack, The Thomas Mitchell Professor of Cognitive Neuroscience, Director of Trinity College Institute of Neuroscience, added: Artificial neural networks were partly inspired by the brain. Much like infants, they rely on learning, but current implementations differ greatly from human (and animal) learning. Through interdisciplinary research, babies can help unlock the next generation of AI.”

Relation: Zaadnoordijk L, Besold TR, Cusack R. Lessons from infant learning for unsupervised machine learning. Nat Mach Intel. 2022;4(6):510-520. doi: 10.1038/s42256-022-00488-2

This article was republished from the following materials. Note: The material may have been edited for length and content. For more information, please refer to the given source.


Comments are closed.