No label defined (Q3700864)

From MaRDI portal

!

This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:

scientific article; zbMATH DE number 3926278
Language Label Description Also known as
default for all languages
No label defined
    English
    No label defined
    scientific article; zbMATH DE number 3926278

      Statements

      1985
      0 references
      Artificial intelligence
      0 references
      Proceedings
      0 references
      Conference
      0 references
      IJCAI
      0 references
      Los Angeles/California
      0 references
      Education
      0 references
      Architectures
      0 references
      hardware
      0 references
      Automated Reasoning
      0 references
      non-monotonic logic
      0 references
      Automatic Programming
      0 references
      Cognitive Modelling
      0 references
      Expert Systems
      0 references
      Knowledge Representation
      0 references
      Learning
      0 references
      Acquisition
      0 references
      Problem solving
      0 references
      Logic Programming
      0 references
      Prolog extensions
      0 references
      Natural Language
      0 references
      Perception
      0 references
      computational geometry
      0 references
      Philosophical Foundations
      0 references
      Planning
      0 references
      Search
      0 references
      Robotics
      0 references
      Theorem Proving
      0 references
      The Proceedings contain in two volumes 263 selected papers within 1309 pages. There are 16 sections of different dimensions. The most extensive sections concern Expert Systems, Natural Language, Learning and Acquisition. The automation of reasoning and programming is met at least in four substantial sections (Automated Reasoning, Theorem Proving, Automatic Programming, Logic Programming) being, quantitatively, the mostly worked out topic in AI. Special emphasis is on Knowledge Representation (one of the ''hardest'' sections by its very interesting and various logical approaches), Cognitive Modelling, Learning and Acquisition. Let us now mention (some of) the topics in the sections. ''AI and Education'' (6 papers) presents experimental and working systems for teaching high school geometry, undergraduate mathematics, learning and debugging pure LISP, second language learning. The section on ''AI Architectures'' (8 papers) contains hardware approaches to the future computers: AND/OR parallelism applicable to Horn-clause logic programs, ultra-concurrent symbolic multiprocessors, many parallel algorithms, the NON-VON massively parallel machine (Columbia University), parallelism illustrated within the Cellular Array Processor (ITT Advanced Technology Center). ''Automated Reasoning'' (22 papers) deals with: counterfactuals described in terms of existing logical operators (Mathew L. Ginsberg), non- monotonic logic as self-choice or self-determination rather than self- knowledge (Jon Doyle), functional approach to non-monotonic logic (Erik Sandewall), quantitative (probabilistic) approach in non-monotonic reasoning. Many papers are concerned with uncertainty inference rules, different non-monotonic systems (default logic, circumscription), propositional dynamic logic of derivations opposed to non-monotonic logic in reason-maintenance systems (Allen L. Brown jun.), non-monotonic process logic implemented in LISP (James W. Goodwin), problem solving using non-chronological backtracking. ''Automatic Programming'' (7 papers), despite its smaller dimension, contains interesting approaches: Knuth-Bendix completion used to synthesize logic and functional programs (Nachum Dershowitz), program inversion in pure LISP (David Eppstein), symbolic execution in algorithm design (David Steier, Elaine Kant), program synthesis for binary-search (Zohar Manna, Robert Waldinger). In the section on ''Cognitive Modelling'' (12 papers) the problems of neuronal models (David Touretzky, Geoffrey Hinton), parallelism in associative memory, computational neurolinguistics (Helen M. Gigley), and input-expectation discrepancy reduction underlying the intelligent behaviour systems are discussed. ''Expert Systems'' (28 papers) presents a great diversity of high quality AI applications. There are analysed not only concrete achievements in the field but also the basic problems of expert system design. It is impossible to review the characteristics of so many realizations but we have to remark the parallel approaches to inference and reasoning, many implementations of Dempster-Shafer evidential-reasoning theories and, in general, uncertainty as efficient inference rules. Many expert systems are designed for medical decision making, (VLSI) circuit design, etc. The ''Knowledge Representation'' section (21 papers) is full of logical models: (decidable) first-order logics for knowledge representation (Peter Patel-Schneider), results on default logic (Witold Lukaszewicz), modal logics of knowledge and belief (Joseph Halpern, Yoram Moses, Ronald Fagin), analysis of monotone knowledge (Moshe Y. Vardi), event calculus, a common-sense theory of time (James F. Allen, Patrick J. Hayes), hybrid reasoning applied to first-order theorem proving (Ronald Brachman, Victoria Gilbert, Hector Levesque), etc. The section on ''Learning and Acquisition'' (31 papers) discusses the quantitative discovery, goal regression as learning search heuristics, the problem of ''what to learn'' not only ''how to learn'', inductive learning, verification-based learning. Problem solving strategies are applied on a large scale in design and implementation of learning by discovery, the use of own performance through learning, learning by observation (concept formation) as opposed to learning from examples (concept identification), higher level learning as a process of categorization. The area of application systems is quite diverse. ''Logic Programming'' (12 papers) presents many Prolog extensions: incorporating many-valued (or fuzzy) logic, extending Prolog structures and inference types, using more efficient unification procedures, providing better tracing and debugging Prolog tools. The efforts can be integrated in a more general trend of ''bridging the gap between logic programming and theorem proving'' (Wolfgang Schönfeld): search strategies for logic programs, Prolog viewed as special Gentzen-type procedure, new techniques of attaching proofs to Skolem functions (Shigeki Goto), verification systems for recursive quantificational logic, etc. In the section on ''Natural Language'' (28 papers) we remark the following main trends: new finite state automata (Glenn David Blank's register vector grammars, Masaru Tomita's graph-structured stacks), logic grammars used in parsing (Fred Popowich), different parallel models (Kurt Eiselt). The macro-linguistic structures are met within discourse computational modelling (Barbara Grosz, Uwe Reyle, and many others), language generation by integrating text planning and text production (Edward Hovy), description-directed natural language generation (David McDonald, James Pustejowsky), text summarization (Danilo Fum, Giovanni Guida, Carlo Tasso), natural language interfaces to data or knowledge bases. ''Perception'' (24 papers) describes many 3-D and 2-D motion or shape analyses using very different approaches (Lie algebras, projective geometry, patterns, stereo paradigm, extended Gaussian, etc.), new algorithms having the roots in computational geometry but using efficient strategies (e.g. parallelism). As in all the other sections, many already working systems are presented. ''Philosophical Foundations'' contains two very interesting papers: whether machines really understand or not is ultimately an ethical question, not a factual one, requiring not an answer but a practical decision (Aaron Sloman). Peter Cheeseman argues ''In Defense of Probability''. From the section on ''Planning and Search'' (20 papers) we mention: mathematical foundations of nonlinear planning, more efficient approaches to the A, \(A^*\), \(SSS^*\) search algorithms, non-monotonic, temporal and automatic planning as applications to reasoning and robot systems, interactions of constraint-satisfaction problems with search techniques. ''Robotics'' (11 papers) have as general characteristics the great integration of vision, tactile sensing, planning and reasoning in concrete robots, the effort to fill the gap between task planning and low level robot languages. There are presented autonomous road-following robots, the HILARE robot project (CNRS - France), Utah/MIT dexterous hand, etc. In ''Theorem Proving'' (13 papers) various methods can be observed generalizing or making more efficient the known strategies: paramodulation for equality relation (Younghwan Lim, Lawrence Henschen), term rewriting and polynomial simplification methods, many sorted calculus with polymorphic functions, more efficient unification algorithms (Manfred Schmidt-Schauss), parallel unification, theory resolution (Mark Stickel), path resolution (Neil Murray, Erik Rosenthal), hierarchical deduction (Tie Cheng Wang) and structural induction (Jacqueline Castaing) theorem proving. The ''General'' section (18 papers) is one of the most exciting: invited papers and panel discussions. It contains brilliant names and useful conclusions, debating topics like: Discovery and Reasoning in Mathematics (Alan Bundy), Building a Bridge Between AI and Robotics (Hirochika Inoue), Self-Knowledge and Self-Representation (John Perry), Understanding and Automating Algorithm Design (Elaine Kant). The panels discussed about AI and the Law (Robert Kowalsky, Marek Sergot, Yorick Wilks, and others), Raising the Standards of AI products, Reasoning with Uncertainty for Expert Systems (Ronald R. Yager), User Modelling, The US and Japan Cooperative AI Research and Development (Howard E. Jacobson), Expert Systems: How Far Can They Go? (Terry Winograd, Randall Davis, Stuart Dreyfus, Brian Smith).
      0 references

      Identifiers

      0 references
      0 references
      0 references
      0 references
      0 references
      0 references
      0 references
      0 references
      0 references
      0 references
      0 references
      0 references