​
Adam Wessell, Editor
PHY2895- Machine Learning & Neural Networks
Spring 2019
Dr. Hawley
​
Best viewed with Chrome browser on Desktop
Brought to you by:
Norbert Wiener
By: Nick Hankins
Wiener was an internationally recognized mathematician born in Columbia, Missouri in 1894. He would go on to win the Bocher Memorial prize in 1933, the National Medal of Science in 1963, and many book awards. As a child prodigy, he graduated with a degree in mathematics from Tufts College when he was 14. With a dissertation focused on math logic, he would go on to get his Ph.D. at Harvard when he was 18.[1] Wiener continued to study math, traveling to Europe to do so. When he returned to the US, WWI had just started. He attempted to enlist, but was denied because of his poor eyesight. At this moment, Wiener decided to explore many different career opportunities. He taught philosophy courses at Harvard, worked some for the General Electric Company, and even became a staff writer for Encyclopedia Americana.[2]
During WWII, Wiener spent most of his time on the automatic firing of anti-aircraft guns. It was this work that inspired one of Wiener’s most impactful contributions to the field of AI: cybernetics. Cybernetics is the science behind automatic control systems and communications in machines and living things. Related to this field, Wiener was the first to formalize the concept of feedback – where previous outputs are routed back into a system as inputs. Wiener said, “The nervous system and the automatic machine are fundamentally alike in that they are devices, which make decisions on the basis of decisions they made in the past.”[3] To gain a better understanding of the decision making and cybernetics in machines, Wiener spent time learning about cognitive science within living things. Additionally, since cybernetics is so closely related to robotics and automated machines, Wiener often crafted ideas and theories about AI. In 1948, Wiener published his book titled
Cybernetics: Or Control and Communication in the Animal and the Machine.
Wiener’s legacy has deep roots within the field of machine learning and AI. His work in cybernetics inspired many mathematicians, engineers, and computer scientists to focus on machine learning and the role of automated machines in everyday life. However, there seems to be a disconnect between the conventional studies of machine learning, and Wiener’s cybernetics. Some believe that current AI applications focus too heavily on fixed assumptions, constant data storage, and the simplistic ability to manipulate data, where information is independent of outside factors. Implementing the ideas of cybernetics would inject a more environmentally holistic focus into AI. The point of view would change from one of objectivity to subjectivity, where the contextual complexity is taken into account. Deep learning is something that is more in line to cybernetics, which makes sense, since there are both derived from the study of biology.
​
Wiener was ahead of his time in the sense that he warned of the danger of putting full trust in automated machines, which can be seen in his book The Human Use of Human Beings from 1950. Wiener believed that the danger was generated from the machine not having the ability to identify with human values; human values which are not always utilitarian.[4]
​
Wiener also has many things in mathematics named after him. To name a few – The Wiener filter (related to information theory), Wiener space (related to measure theory), and the Wiener equation (related to stochastic fluid motion).
​
Norbert Wiener’s work as a mathematician and philosopher in the fields of cybernetics and AI has proven to be both foundational and inspiring for machine learning researchers and thinkers in the field today.
​
Source: ams.org
References
[1] The Editors of Encyclopaedia Britannica. "Norbert Wiener." Encyclopædia Britannica. November 22, 2018
​
[2] O'Connor, J. J., and E. F. Robertson. "Norbert Wiener." MacTutor History of Mathematics Archive.
​
[3] West, Doug. "Norbert Wiener, the Father of Cybernetics." Owlcation. December 08, 2018.
​
[4] Perez, Carlos E. "What Deep Learning Can Learn from Cybernetics." Medium.com. December 16, 2018.