Computers can win at chess, but when solving complex problems — like directing a robot to play catch — computers can't perform as well as a 5-year-old. A new programming paradigm called Hierarchical Temporal Memory (HTM) applies hierarchical memory nodes to increasingly complex information, allowing for information and learning reuse that appears to exceed previous attempts using neural networks. HTM is both a biological and a mathematical model. HTMs work best when there is hierarchical structure in the data — as found in a business - layers of managers, supervisors, and workers, as well as functions like finance and accounts receivables.
The preceding article is a "sneak peek" from Plant Engineering, a newsletter from GlobalSpec. To stay up-to-date and informed on industry trends, products, and technologies, subscribe to Plant Engineering today.