Co-founded the MIT Artificial Intelligence Laboratory and helped make AI an organized research program.
Clarify a term, inspect evidence, or place an idea in context.
inAiLoading page content...
Clarify a term, inspect evidence, or place an idea in context.
inAiLoading page content...
Profile · symbolic and plural AI
Plan for this page
Minsky helped establish major AI institutions and explored intelligence as an interaction among many specialized processes rather than one simple mechanism.
Contributions
Co-founded the MIT Artificial Intelligence Laboratory and helped make AI an organized research program.
Surveyed early AI methods in 'Steps Toward Artificial Intelligence,' linking search, pattern recognition, learning and planning.
Developed the Society of Mind view that intelligence may emerge from coordination among many limited processes or agents.
Contributed to early neural-network experiments, robotics, machine perception and knowledge representation.
Minsky's career crossed approaches that are now often presented as opposites. He built an early neural-network learning machine, developed symbolic and common-sense theories, and later argued that intelligent systems would need multiple kinds of representation rather than one universally correct method.
With Seymour Papert, Minsky analyzed capabilities and limitations of particular perceptron architectures. That mathematical work became part of a larger historical shift in research funding and fashion, but claims that the book single-handedly ended neural-network research are too simple and omit technical, institutional and economic factors.
Context, limits, and debates
Sources and scope
MIT account of Minsky's institutional and technical work, including the AI Laboratory, robotics and theories of mind; it is a commemorative source and not a neutral evaluation of every claim.
Primary 1961 survey setting out early AI problems and methods across search, learning, pattern recognition and planning; reflects the field and expectations of its time.
Primary essay arguing for heterogeneous architectures rather than a single symbolic or connectionist method; useful for correcting simplified accounts of Minsky's position.
Publisher description of Minsky and Seymour Papert's analysis of a class of perceptrons and its later expanded edition; it does not support the simplistic claim that one book alone caused an 'AI winter.'