How to improve computer’s learning ability
Computer-based AI (AI) can function more like human intelligence when programmed to use a way faster technique for learning new objects, researchers say.
In the journal Frontiers in Computational Neuroscience, the researchers explained how the new approach vastly improves the power of AI software to quickly learn new visual concepts.
“Our model provides a biologically plausible way for artificial neural networks to find out new visual concepts from alittle number of examples,” said the researcher, Maximilian Riesenhuber, Professor of neuroscience at Georgetown University within the US.
“We can get computers to find out far better from few examples by leveraging prior learning during a way that we expect mirrors what the brain is doing,” Riesenhuber added.
Humans can quickly and accurately learn new visual concepts from sparse data — sometimes one example.
Even three- to four-month-old babies can easily learn to acknowledge zebras and distinguish them from cats, horses, and giraffes. But computers typically got to “see” many samples of an equivalent object to understand what it’s , the researcher explained.
The big change needed was in designing software to spot relationships between entire visual categories, rather than trying the more standard approach of identifying an object using only low-level and intermediate information, like shape and color, he added.
The researchers found that artificial neural networks, which represent objects in terms of previously learned concepts, learned new visual concepts significantly faster.
The brain architecture underlying human visual concept learning builds on the neural networks involved in visual perception .
The anterior lobe of the brain is assumed to contain “abstract” concept representations that transcend shape. These complex neural hierarchies for visual recognition allow humans to find out new tasks and, crucially, leverage prior learning.
“Our findings not only suggest techniques that would help computers learn more quickly and efficiently, they will also cause improved neuroscience experiments aimed toward understanding how people learn so quickly, which isn’t yet well understood,” Riesenhuber concluded.