The cortex in the code: how neuroscience makes AI intelligent
Abstract:
Today’s AI algorithms closely parallel the function of the brain, such that they are able to do many tasks as well as, or better than, humans. This talk will examine not only the functional parallels between the brain and AI but also the structural parallels, drawing comparisons between AI and the brain to discuss AI design in a creative way.
Based on the similarities between the two, there are a few good questions to ask:
- How do the building blocks of the brain, such as neurones and synapses, create thought?
- What are the building blocks of AI, how are they similar to the brain, and how do they create intelligence?
- In the future, how can we use concepts from neuroscience to develop better algorithms?
- Can we use neuroscience to develop super-intelligent AI?
In this talk, we’ll examine the construction and foundation of various algorithms. We’ll also dive into how these foundations impact function in both the brain and AI. Information processing, prediction, representation, and entropy, from the standpoint of both the brain and AI, will be introduced. We’ll also discuss how neuroscience will continue to inspire AI. By the end of this talk you will have a unique understanding of the way AI algorithms are built.
This talk will serve as a great introduction to AI algorithms, as well as a fresh perspective for those more familiar with AI. No knowledge of AI nor neuroscience is required.
Talk: link to recording
Slides: PyData 2019.pdf
Other details: link to website