scott@bgsu.CSNET (Tom Scott) (06/19/86)
The moderator of the AI-List has made an impassioned plea for help. I would like to help, but before I offer to start a new Arpanet, Csnet, or UUCP newsgroup, I'd like to put forth an organization to Ken's list of possible new newsgroups. This organization comes from the Japanese side of the Pacific, and is outlined by Brian Gaines in a recent article, "Sixth Generation Computing: A Conspectus of the Japanese Proposals" ("SIGART Newsletter", January 1986, pp. 39-44). Figure 1 of the article, complemented by the fundamental topics that I've added for the sake of completeness, cuts the cake thusly: Theoria | Praxis | Techne ------------ | -------------------- | -------------------- | Expert systems | Pattern recognition Physiology | | Cognition | Machine translation | Learning Psychology | systems | Problem solving | | Natural language Linguistics | Intelligent CAD/CAM | Image processing | systems | Speech recognition Logic | | Man-Machine interface | Intelligent robotics | ============================================================= | Managerial | Expert systems Epistemology | cybernetics | | Decision support | Development languages Modern logical| systems | and environments metaphysics | Information | | retrieval systems | Computing/knowledge Vedic Science | | machines ============================================================= THE UNIFIED FIELD OF ALL POSSIBILITIES This is the world of the sixth generation: knowledge science and knowledge systems. The fifth generation, which deals mainly with the daily realities of knowledge engineering and expert systems, as well as with the advanced research and development of VLSI architectures for the processing of Prolog code and database systems, is distinct from the sixth generation. To get a better feel for these distinctions, I'd like to suggest the following homework assignment for new newsgroup moderators: (1) Read Brian's article. (2) Read the abstract of the paper that I'll be presenting to the sixth-generation session at the 1986 International Conference on Systems, Man, and Cybernetics (Atlanta, October 14-17); the abstract is appended to this message. (3) Think before you flame; then write back to me or to this newsgroup and share your thoughts. We are children of the cybernetic revolution and we are witnessing the rising sunshine of the Age of Enlightenment. Tom Scott CSNET: scott@bgsu Dept. of Math. & Stat. ARPANET: scott%bgsu@csnet-relay Bowling Green State Univ. UUCP: cbosgd!osu-eddie!bgsuvax!scott Bowling Green OH 43403-0221 ATT: 419-372-2636 * * * Abstract of the sixth-generation SMC paper * * * KNOWLEDGE SCIENCE The Evolution From Fifth-Generation Expert Systems To Sixth-Generation Knowledge Systems Theory, practice, technology--these are the makings of a full vision of knowledge science and sixth-generation knowledge systems. Prior to the establishment of research and development projects on the Fifth Generation Computing System (FGCS), knowledge science did not exist independent of knowledge engineering, and was conceptualized only in technological terms, namely, expert systems and "machine architectures for knowledge-based systems based on high-speed Prolog and relational database machines" (Gaines 1986). Although the design and development of fifth-generation machines and expert systems will continue for years to come, we want to know now what can be done with these ultra-fast architectures and expert systems. What kinds of knowledge, other than the knowledge of domain experts in fifth-generation expert systems, can be acquired and encoded into sixth-generation knowledge systems? What can be done on top of fifth-generation technology? How can fifth-generation architectures and expert-system techniques be extended to build intelligent sixth-generation knowledge systems? Beyond the fifth generation it is necessary to envision practical applications and theoretical foundations for knowledge science in addition to the technological implementation of machine architectures and expert systems. This paper discusses the full three-part vision of knowledge science (theoria, praxis, and techne) that is emerging around the world and has been treated by the Japanese under the title Sixth Generation Computing System (SGCS). Theoria: As indicated in Brian Gaines's article, "Sixth Generation Computing: A Conspectus of the Japanese Proposals" ("ACM-SIGART Newsletter" January 1986), the theoretical foundations of knowledge science are arranged in levels, proceeding downward from physiology to psychology to linguistics to logic. Continuing in this direction toward deeper foundations, the field of knowledge science embraces epistemology and modern logical metpahysics. On the empirical side of the deep foundations is the probability-based epistemology of pragmatism, explicated in Isaac Levi's "The Enterprise of Knowledge" (1980); on the transcendental side are Immanuel Kant's "Critique of Pure Reason" (1781-87) and Edmund Husserl's "Formal and Transcendental Logic" (1929). A simplified diagram of the four main divisions of mind, based on one sentence of the Critique ("Beide sind entweder rein, oder empirisch": B74), is: Understanding Sensibility | E Knowledge Images m of --------> Objects p objects | | ----------------------+----------------------- T | r Pure concepts Schemas Pure forms of a (categories) --------> intuition n and principles | (space and time) s | Praxis: The SGCS project is also concerned with the practical applications of knowledge science. These applications are organized under four headings: expert systems, machine-translation systems, intelligent CAD/CAM, and intelligent robotics. Another way of organizing the applications of knowledge science in terms familiar to the IEEE Systems, Man, and Cybernetics Society is: managerial cybernetics, organizational analysis, decision support, and information retrieval. Stafford Beer's "The Heart of Enterprise" (1979) is the focal point of our discussion of knowledge-science praxis. Techne: The SGCS project targets eight technological areas as the basis for the future research and development of sixth-generation knowledge systems: pattern recognition, cognition, learning, problem solving, natural language, image processing, speech recognition, and man-machine interfacing. To fully realize the R&D potential of these eight areas, sixth-generation knowledge scientists must be on friendly terms with the following areas of expertise from fifth-generation knowledge engineering: (1) Expert systems. (a) Concepts and techniques for the acquisition, representation, and use of knowledge. (b) The software engineering of knowledge systems, including a methodology for the building of expert systems and the management of expert-system development teams. (c) Expert systems and shells. (2) Three levels of systems and software. (a) Production systems (e.g., ITP, Prolog, and OPS83). (b) Traditional AI/KE languages (e.g., Lisp and Prolog). (c) Development environments and utilities (e.g., Unix, C, and Emacs). (3) The knowledge engineer's technical intuition of a computational knowledge machine. (a) Lambda Consciousness, based on the idea of a Lisp machine. (b) Relational database machines. (c) Prolog machines. The paper includes observations from the experience of the University of Wisconsin-Green Bay in its attempts to establish a regional knowledge-engineering and knowledge-science resource center in the Northeastern Wisconsin area. * * * Finis * * *