Ideas: 2026

April 2026

• Can we do machine learning with a neural network that has multiple depths? e.g. a 30-layer network that also has pathways that will skip multiple layers? Thinking about this as an analogue of focused vs. diffuse thinking.