Archives of General Psychiatry · 1962 · 511 citations · 0 references
Life SciencesComputer ScienceNeuroscienceArtificial IntelligenceComputational NeuroscienceNeural EngineeringCognitive ModelingNeural Circuit ModelingNeural ModelingNeural NetworksNeural Signal ProcessingNeural System ModelingComputational Neuroscience ModelsNeural Circuit ArchitecturePerceptron
Recent engineering projects model brain activity for pattern recognition by assuming it operates via algorithmic methods similar to digital computers, a premise that Rosenblatt’s perceptron approach challenges by focusing on the physical structures underlying natural intelligence. Critics such as Lashley argue that computer‑simulated behavior is artificial and that such models rely on extrabiological principles, undermining their biological validity.
In recent years, there have been a number of engineering projects concerned with the design of brain models for pattern recognition and artificial intelligence. The basic assumption, underlying these projects is that the brain operates by built-in algorithmic methods similar to those employed in modern digital computers. Hence, nervous activity can be simulated by these computers. The value of such a program has been challenged by Lashley and others on the grounds that computer-simulated behavior is artificial, that the model is an invention operating on extrabiological principles. In this formidable book, Rosenblatt has offered a somewhat different program involving the design and testing of brain models described as perceptrons. His program is concerned not with devices for artificial intelligence, but, rather, with "the physical structures and neurodynamics principles which underly "natural intelligence." A perceptron consists of a set of signal generating units ("neuro-mimes") connected together to form a network.