Perspectives

   

• [Bobrow1993] Artificial Intelligence in Perspective.
  Take a look at this article and browse through this great volume of AIJ.
• [Lenat and Feigenbaum1991] On the thresholds of knowledge.
  A general view with many interesting hypotheses in AI. Note that the physical symbol system hypothesis is wrong as stated.
• [McCarthy1981] Some epistemological problems in AI.
  An old (original version from 1977) but still relevant paper that outlines several representational challenges, many of which still have not been adequately addressed.
• [Brooks1991] Intelligence without representation.
  Brooks argues that intelligent entities can be constructed without making use of explicit representations of knowledge.
• [Guha and Lenat1990] Cyc: a midterm report.
  Computers need common sense to prevent brittleness and for semantic disambiguation. CYC is an attempt to create a large commonsense knowledgebase spanning human knowledge. This report describes some lessons learned over the first five years and looks at plans and expectations for the next five years.

• [Turing1963] Computing Machinery and Intelligence.
  Turing's vision from 1950 is amazing. He foresees logical inference, control of search, and learning as important tasks to tackle.
• [McCarthy1985] Programs with common sense.
  McCarthy outlines his vision (in 1959) for an `Advice Taker' program whose performance could be improved by receiving advice during the course of its actions. He argues for the use of a declarative representation of information in this program.
• [Hayes1985] The second naive physics manifesto.
  An interesting exposé of the difficulties involved in formalizing commonsense reasoning. The paper is somewhat long; skim it for an understanding of the naive physics enterprise.