About this Item
December 15, 1965. 4to (280 × 215mm), pp. 81, [1]. Photomechanically reproduced typescript, staple-bound, some creasing, file number sticker to the front page, else very good. Presentation copy, inscribed by Lederberg on the front page: "I think you will enjoy this. J." Ex Abraham Pais. Part II of Lederberg s DENDRAL-64 series, supported by NASA, the National Science Foundation, and the National Institutes of Health, and submitted from the Department of Genetics, School of Medicine, Stanford University. Where Part I established a canonical formulation for chemical graphs that are pure trees (that is, acyclic molecular structures representable as single branching hierarchies), Part II addresses the harder problem of ring structures, defined here as strictly cyclic graphs: sets of atoms not separable by fewer than two cuts. Part III, as the introduction indicates, will extend the analysis to complete structures in which rings appear as special nodes within a larger tree. The report is concerned with the topology of cyclic graphs as a foundation for the systematic computer enumeration of organic ring systems, from simple cycles such as benzene to polycyclic structures including the steroid nucleus, the morphine nucleus, naphthalene, and biphenyl, each assigned a unique graph-theoretic encoding. The objectives set out for the treatment of trivalent cyclic graphs (enumeration of all possible graphs, identification of isomorphisms between superficially different representations, rational description and ordering of each graph, characterization of symmetries, rational numbering of vertices and paths, and compact computable notation) define a program that is recognizably that of modern cheminformatics, pursued here at the moment of its inception. This report is an early technical foundation of the DENDRAL project, the first expert system in the history of artificial intelligence, which sought to encode the reasoning of a chemist skilled in mass spectrometric structure determination and make it executable by machine. Lederberg s graph-theoretic approach to molecular structure was the mathematical substrate on which that ambition rested. The methods developed here thus stand at the origin of both computational chemistry and machine reasoning about scientific problems, two fields whose subsequent development has touched virtually every area of drug discovery, materials science, and chemical research. The recipient, Abraham Pais, was theoretical physicist, historian of twentieth-century science, and author of the definitive scientific biographies of Einstein and Bohr; a provenance that unites two of the most consequential scientific intellects of the postwar decades. In the context of artificial intelligence, DENDRAL-64 Part II is best understood not as a curiosity of chemical graph theory but as an essential structural pillar of the first expert system, and the one that set the template for all that followed. What Lederberg accomplishes here is to make molecular structures (especially the complex ring systems that had resisted systematic treatment) available to symbolic reasoning, giving the computer a precise, unambiguous language in which to generate, label, and compare them. Without that foundation, the larger DENDRAL ambition was simply impossible: "chemists intuition" about plausible structures cannot be translated into executable rules unless the structures themselves can be represented without ambiguity or redundancy. With it, DENDRAL could do something no program had done before: systematically explore the space of chemically possible candidates, discard equivalents, and converge on solutions that matched experimental data, behaving, in effect, like a genuine domain expert. In this sense, the report stands at the effective origin point of knowledge?based AI in scientific practice, setting the template for how later systems in medicine and engineering would encode and manipulate expert knowledge. While its techniques belong to th.
Seller Inventory # 1299
Contact seller
Report this item
