Designing Autonomous AI

Chapter 3. How Brains Learn Best: Teaching Humans and AI

Recent research suggests that the brain builds structures on the fly in 11 dimensions. Apparently, the brain assembles neurons into structures of various shapes and complexity as it processes information and makes decisions. There seems to be a link between learning and the design of structured intelligence. This matches my experience with brain design, and it matches what we know about teaching. There is intelligence in the design of the human brain (the arrangement of the neurons), and that design is dynamic based on the thought process. Similarly, there is intelligence in the design of AI brains, and better designs learn better. This chapter lays the groundwork for what a brain design is and how to use one to guide learning.

Learning Multiple Skills Simultaneously Is Hard for Humans and AI

I hope that you agree at this point that complex tasks often require multiple skills and strategies, and that those skills and strategies are likely dictated by the dynamics of the task itself. Chess strategies work within and because of the game’s own rules. Basketball strategies work within and because of the dynamics of basketball. Here’s the problem: it’s proven quite difficult for humans and AI brains to practice multiple skills simultaneously.

The “Ray Interference” paper, published by Google DeepMind in 2019, asserts and empirically proves that AI brains that store learning in neural networks get confused and take a long time ...