Allen Newell

• 1. From Physics To Organization

  ›

  Allen Newell stood at Stanford University in 1949 holding a bachelor's degree in physics. He moved to Princeton University that same year to study mathematics as a graduate student. The field of game theory appeared before him during these early academic years. It convinced him that pure mathematics lacked the experimental component he desired. He sought a path combining theoretical work with hands-on research instead. In 1950, Newell left Princeton for Santa Monica California. He joined the RAND Corporation where a group studied Air Force logistics problems. Joseph Kruskal worked alongside him on theories about organizational structures. They created A Model for Organization Theory and Formulating Precise Concepts in Organization Theory. Newell eventually earned his PhD from Carnegie Mellon under Herbert Simon. His focus shifted toward laboratory experiments on decision making within small groups.

  12

• 2. The Logic Of Machines

  ›

  September 1954 marked a turning point when Oliver Selfridge described a computer program recognizing letters. Newell believed systems could contain intelligence and adapt to new situations. He wrote The Chess Machine: An Example of Dealing with a Complex Task by Adaptation in 1955. This document outlined an imaginative design for playing chess like a human. Herbert A. Simon noticed this work and partnered with programmer J. C. Shaw. Together they developed the Logic Theorist program in 1956. It became one of the earliest true artificial intelligence programs ever created. Newell introduced list processing as a programming paradigm used by AI since then. He applied means-ends analysis to general reasoning or what he called reasoning as search. He also utilized heuristics to limit the search space during problem solving. These inventions laid the foundations for the entire field of artificial intelligence research.

  4517

• 3. Building Soar And Cognition

  ›

  Newell and Simon formed a lasting partnership at Carnegie Mellon University. They founded an artificial intelligence laboratory there in the late fifties. Their work included the General Problem Solver released in 1957. This implementation featured highly influential means-end analysis techniques. They proposed the physical symbol systems hypothesis asserting all intelligent behavior reduces to symbol manipulation. Newell continued refining these ideas until his death on the 19th of July 1992. His final efforts focused on developing a cognitive architecture known as Soar. He published a unified theory of cognition in 1990. The field of cognitive architectures remains active within both AI and computational cognitive science communities today. One of his last letters discussed ongoing improvements to this theoretical framework.

  517

• 4. The Dartmouth Gathering

  ›

  Researchers gathered informally at the Dartmouth conference in 1956. Newell presented the Logic Theorist program alongside Herbert A. Simon and J. C. Shaw. This event is now widely considered the birth of artificial intelligence as a distinct field. Those who attended became leaders of AI research for the next two decades. Newell was among those who shaped the direction of the discipline. The conference brought together scientists interested in simulating intelligence with machines. It established a community that would drive innovation throughout the following years. The gathering transformed scattered experiments into a cohesive area of academic study.

  417