Systematic Methods of Scientific Discovery
Papers from the AAAI Spring Symposium
Scientific discovery is among the most celebrated creative processes, and receives scholarly attention from several disciplines including AI. Experience has shown that some scientific reasoning can be made systematic. This symposium will examine the state of systematic methods for scientific inference, with special emphasis on insights that promise broad applicability within science. The development of such methods will, as foreseen by Allen Newell, increasingly lead to the sciences taking a metaposition, in which doing science will involve understanding the information processes of scientific inference, and building systems that do the object-level science, leading to a convergence of AI and the scientific enterprise. This symposium will feature: descriptions of existing methods or systems and their achievements in natural, social, and mathematical science; generalizations made over these systems; exploratory task analyses; and psychological/historical studies that promise to lead to further systematic methods.