Hierarchical Temporal Memory

In hierarchical temporal memory, researchers attempt to understand how the brain works and learn how to produce models and simulations of neocortex behavior.

Hierarchical temporal memory has been called a "new approach to artificial intelligence," which focuses on universal learning and not on task-specific processes. For example, part of the hierarchical temporal memory principle involves "deep learning" algorithms with templates at various layers, where scientists can evaluate how the human brain perceives images and puts them together from various components.

In interviews, Hawkins has argued that the idea of universal learning algorithms is going to be the dominant paradigm and that this will enhance many areas of the tech industry. Part of the theory is that the neocortex uses a single framework for disparate items like vision, motor activity, planning and much more. Scientists have also found that two different types of pattern recognition should be put together to model the behavior of the brain. All of these will lead to a more robust way to create systems that mimic the higher level functions of the brain.