You might not know the name Demis Hassabis, but you're most likely familiar with his work. Known as "Google's Intelligence Designer,"
Hassabis founded DeepMind, the London-based leader in artificial intelligence (AI) that was later acquired by Google. From AlphaGo
to broadening all of machine learning
, Hassabis is at the forefront of progress and innovation in AI. So the industry was certainly listening when Hassabis recently stated
that only through pushing forward with neuroscience and a better understanding of actual human intelligence can we hope to realize the full potential of AI. This makes a lot of sense—if AI is moving away from basic machine learning
and into the realm of functioning like a real "intelligence," then we should probably have a fairly complete understanding of human intelligence itself. To expand on this idea, Hassabis and three of his colleagues recently published a paper in Neuron
Image Credit: aytuguluturk / pixabay
The paper focuses on two aspects of this idea. First, they assert that a better understanding of how the human brain functions will allow for further progress in the development of new AI tech and algorithms. Secondly, the development of these innovative AIs could, reversely, help researchers to better understand and define intelligence. Here's how they put it:
The benefits to developing AI of closely examining biological intelligence are two-fold. First, neuroscience provides a rich source of inspiration for new types of algorithms and architectures, independent of and complementary to the mathematical and logic-based methods and ideas that have largely dominated traditional approaches to AI. For example, were a new facet of biological computation found to be critical to supporting a cognitive function, then we would consider it an excellent candidate for incorporation into artificial systems. Second, neuroscience can provide validation of AI techniques that already exist. If a known algorithm is subsequently found to be implemented in the brain, then that is strong support for its plausibility as an integral component of an overall general intelligence system.
So integrating neuroscience research with AI development will not only allow AI to advance, it will also improve our understanding of intelligence itself. It's a win-win situation where both fields aid each other in a very natural way. The paper shows how, while the two seem a likely pair, biology has not been as heavily considered in AI research as it should be. There has been a lot of discussion as to what "true AI" really is, and whether or not it is possible (disregarding the ethical issues) to create AI that is a realistic replica of human intelligence. Perhaps, given Hassabis' expertise and leverage within the field, AI will actually start shifting away from basic machine learning
and towards more complex forms of "intelligence."