Researchers at a Scottish university say machine learning developers need to understand better whether AI processes information like a human.
An article in the Trends in Cognitive Sciences journal based on research from University of Glasgow’s School of Psychology and Neuroscience questions whether the human brain and its Deep Neural Network (DNN) models recognise things in the same way, using similar steps of computations.
DNNs are increasingly used in everyday applications such as automated face recognition systems and self-driving cars, while researchers use them to model the processing of information. Despite its success, scientists admit they still do not fully understand how DNNs work.
The university’s research technology dean, Philippe Schyns, says a better understanding would allow for more accurate real-world applications. “If we have a greater understanding of the mechanisms of recognition in human brains, we can then transfer that knowledge to DNNs, which in turn will help improve the way DNNs are used in applications such as facial recognition, where they are currently not always accurate,” he says.
“Creating human-like AI is about more than mimicking human behaviour – technology must also be able to process information, or ‘think’, like or better than humans if it is to be fully relied upon.
“We want to make sure AI models are using the same process to recognise things as a human would, so we don’t just have the illusion that the system is working.”
The study, ‘Degrees of Algorithmic Equivalence between the Brain and its DNN Models’ is published in Trends in Cognitive Sciences. The work is funded by Wellcome and Physical Sciences Research Council.