What is Machine Intelligence?

And how does it differ from related terms like Artificial Intelligence and Machine Learning?

When we think about the progress of computing, we often hear three phrases come to the forefront: Machine Learning, Machine Intelligence, and Artificial Intelligence. Many believe these three phrases stand in for each other and refer to the same thing. They do not. Additionally, automation throws another word into the mix. In short, automation consists of a machine doing exactly what a human being tells it to do. Exactly the same. Every time. Even if it’s wrong. We see automation deployed in many industries today, and it often serves well. Still, automation does not approach machine learning, or machine intelligence, or artificial intelligence. When we think about and deploy machine learning, we want machines to look for and recognize tendencies or trends within the data it analyzes. As the machine “learns” from its analysis of the data, it writes new ways of acting while extending functions or avoiding pitfalls that its analyses taught it. Thus, accuracy and efficiency increase without human intervention at each juncture of analysis. When we talk about machine intelligence, we generally think of machines programmed in ways to use deductive reasoning, e.g., starting with a hypothesis and examining possibilities to reach a logical conclusion. We may also assume that machine intelligence allows machines to recognize mistakes it might make during deductive reasoning and change responses or actions so it does not repeat those mistakes. "In deductive inference, we hold a theory and based on it we make a prediction of its consequences. That is, we predict what the observations should be if the theory were correct. We go from the general — the theory — to the specific — the observations," said Dr. Sylvia Wassertheil- Smoller, a professor emerita at Albert Einstein College of Medicine. This defines the intent of machine intelligence. As machine intelligence acts, we expect it to access multiple machine learning operations while firing them in the correct order to reach the desired goal, and these actions help define machine intelligence and serve as the next level up from machine learning. Finally, artificial intelligence makes sense of all the data ingested and actions taken by machine learning and machine intelligence, and many predict that it may supersede the human mind. As Jeremy Achin put it at Japan AI Experience in 2017, "AI is a computer system able to perform tasks that ordinarily require human intelligence... Many of these artificial intelligence systems are powered by machine learning, some of them are powered by deep learning and some of them are powered by very boring things like rules." Without doubt, if “thinking like humans” comes true, it will require the assistance of machine learning and machine intelligence.

Back