Presenter
Zan Huang
Zan Huang is a researcher with a passion for alternative computational models in artificial intelligence, mass social patterns, chaotic and emergent systems, and linguistics. Currently focused on scaling deep neural networks through neurologically inspired modularity, he explores critical questions around reducing parameter space, enhancing interpretability, and developing self-similar task divisions akin to brain functionality. In 2024, Zan was honored as a Foresight Institute Fellow in Neurotechnology, a recognition highlighted in a press release on the Foresight Institute's website. Additionally, he is the Co-Founder of Myelin Neurotech, a nonprofit dedicated to advancing neurotechnology research. More information about his work can be found on the Myelin Neurotech page.
Summary:
Why is the study of neuroscience still relevant to AI? I take the strong hypothesis that not only is it relevant, but also serves as an essential component in understanding intelligence. 1. Certain physical and mathematical laws scale at multiple levels beyond their original application, and biology is not an exemption from these considerations. 2. The human brain is an example of such a system. 3. The principles that have worked well in deep learning and modern AI point to a particular style of organization not unlike if not identical to how the brain operates. I will be giving an argument for these ideas through the lens of the ontological bases for science, modern deep learning and neuroscientific research, and the path my current research is taking.
