Pedro Domingos Views Neural Networks as Embedded Symbolic Systems
The computer science professor emeritus posted his perspective on the nature of neural networks.
TLDR
Pedro Domingos, Professor Emeritus of Computer Science at the University of Washington and author of The Master Algorithm, made a statement in a public post. He asserted that neural networks function as symbolic systems. According to his description, symbols are embedded inside the networks while rules possess weights attached to them. The message positions this as a way to understand the structure of neural networks. The post originates from his account and reflects his expertise as an academic focused on artificial intelligence topics. No additional context or external confirmation is provided in the packet regarding reactions or further elaboration.
Combined views
8.4K
1 Source, first seen 25d ago