Mark I Perceptron
In plain wordsA room-sized machine built around 1960 that learned to recognize simple shapes and letters through a camera made of 400 light sensors. It was one of the first machines built to learn from examples.
A custom electromechanical computer built at the Cornell Aeronautical Laboratory in Buffalo, New York, to implement Frank RosenblattFrank Rosenblatt, prop. n.: An American psychologist who invented the perceptron, the first neural network that learned from examples, and built a machine that learned to recognize simple images.'s perceptronperceptron, n.: One of the first learning machines: a single artificial neuron that sorts inputs into two groups by drawing a straight dividing line, nudging its weights whenever it guesses wrong. for image recognition. Its 'retina' was a 20 × 20 array of 400 cadmium-sulfide photocells, wired at random through a plugboard to 512 association units whose outputs fed eight response units. The adjustable weights were motor-driven potentiometers that the perceptron learning rule turned during training. Shown printed letters and simple shapes, the machine learned to classify them after repeated presentations and corrections. Its success, and the publicity around it, fueled the first wave of neural-network research; its limits became a target of Minsky and Papert's PerceptronsPerceptrons (book), prop. n.: A 1969 book by Marvin Minsky and Seymour Papert proving that simple one-layer neural networks cannot learn certain patterns. It is widely blamed for a decade-long slump in neural-network research. (1969).
Rosenblatt simulated the perceptron on an IBM 704 before the Mark I Perceptron existed in hardware.
Origin
Funded by the US Office of Naval Research. The perceptron was first simulated on an IBM 704 in 1958, when a press briefing produced extravagant newspaper forecasts of machines that would walk, talk and be aware of themselves; the Mark I hardware was completed and publicly demonstrated in 1960. It is preserved at the Smithsonian's National Museum of American History.
Related
- perceptronperceptron, n.: One of the first learning machines: a single artificial neuron that sorts inputs into two groups by drawing a straight dividing line, nudging its weights whenever it guesses wrong.
- Frank RosenblattFrank Rosenblatt, prop. n.: An American psychologist who invented the perceptron, the first neural network that learned from examples, and built a machine that learned to recognize simple images.
- Perceptrons (book)Perceptrons (book), prop. n.: A 1969 book by Marvin Minsky and Seymour Papert proving that simple one-layer neural networks cannot learn certain patterns. It is widely blamed for a decade-long slump in neural-network research.
- connectionismconnectionism, n.: The view that minds and intelligent machines are best built from huge networks of simple, neuron-like units whose connection strengths are learned, rather than from explicit rules and symbols.
- ADALINEADALINE, prop. n.: An early learning machine from 1960 that adjusted its settings to shrink its errors step by step, the same basic idea used to train today's neural networks.
- neural networkneural network, n.: A kind of computer program that learns a task from examples by tuning thousands to billions of connections between simple number-crunching units, an arrangement loosely inspired by nerve cells in the brain.
Sources
- Rosenblatt, F. (1958). The Perceptron: A Probabilistic Model for Information Storage and Organization in the Brain. Psychological Review 65(6), 386–408.
Mentioned in
Frank RosenblattFrank Rosenblatt, prop. n.: An American psychologist who invented the perceptron, the first neural network that learned from examples, and built a machine that learned to recognize simple images., perceptronperceptron, n.: One of the first learning machines: a single artificial neuron that sorts inputs into two groups by drawing a straight dividing line, nudging its weights whenever it guesses wrong., Perceptrons (book)Perceptrons (book), prop. n.: A 1969 book by Marvin Minsky and Seymour Papert proving that simple one-layer neural networks cannot learn certain patterns. It is widely blamed for a decade-long slump in neural-network research., ADALINEADALINE, prop. n.: An early learning machine from 1960 that adjusted its settings to shrink its errors step by step, the same basic idea used to train today's neural networks.