Keynote Addresses

 2. Professor Xin Yao

Precisiated Natural Language -- Toward a Radical Enlargement of the Role of Natural Languages in Information Processing, Decision and Control

Lotfi A. Zadeh
Professor in the Graduate School and Director, Berkeley Initiative in Soft Computing (BISC), Computer Science Division, Department of EECS, University of California, Berkeley, CA 94720-l776;  Telephone: 5l0-642-4959;  Fax:5l0-642-l7l2;
E-mail:  zadeh@cs.berkeley.edu;
http://www.cs.berkeley.edu/People/Faculty/Homepages/zadeh.html,
Research supported in part by ONR Contract N000l4-99-C-0298, NASA Contract NCC2-l006,
NASA Grant NAC2-ll7, ONR Grant NOOOl4-96-l-0556, ONR Grant FDNOOl499l035, ARO
Grant DAAH 04-96l-034l and the BISC Program of UC Berkeley.

Abstract 

It is a deep-seated tradition in science to view the use of natural languages in scientific theories as a manifestation of mathematical immaturity.

The rationale for this tradition is that natural languages are lacking in precision. However, what is not recognized to the extent that it should, is that adherence to this tradition carries a steep price. In particular, a direct consequence is that existing scientific theories do not have the capability to operate on perception-based information exemplified by "Most Finns are honest." Such information is usually described in a natural language and is intrinsically imprecise, reflecting a fundamental limitation on the cognitive ability of humans to resolve detail and store information. Because of their imprecision, perceptions do not lend themselves to meaning-representation through the use of precise methods based on predicate logic. This is the principal reason why existing scientific theories do not have the capability to operate on perception-based information.

In a related way, the restricted expressive power of predicate-logic-based languages rules out the possibility of defining many basic concepts such as causality, resemblance, smoothness and relevance in realistic terms. In this instance, as in many others, the price of precision is over-idealization and lack of robustness.

In a significant departure from existing methods, in the approach which is described in this talk the high expressive power of natural languages is harnessed by constructing what is called a precisiated natural language (PNL).

In essence, PNL is a subset of a natural language (NL) -- a subset which is equipped with constraint-centered semantics (CSNL) and is translatable into what is called the Generalized Constraint Language (GCL). A concept which has a position of centrality in GCL is that of a generalized constraint expressed as X isr R, where X is the constrained variable, R is the constraining  relation, and isr (pronounced as ezar) is a variable copula in which r is a discrete-valued variable whose value defines the way in which R constrains X. Among the principal types of constraints are possibilistic, veristic, probabilistic, random-set, usuality, and fuzzy-graph constraints.

With these constraints serving as basic building blocks, more complex (composite) constraints may be constructed through the use of a grammar. The collection of composite constraints forms the Generalized Constraint Language (GCL). The semantics of GCL is defined by the rules that govern combination and propagation of generalized constraints. These rules coincide with the rules of inference in fuzzy logic (FL).

A key idea in PNL is that the meaning of a proposition, p, in  PNL may be represented as a generalized constraint which is an element of GCL. Thus, translation of p into GCL is viewed as an explicitation of X, R and r. In this sense, translation is equivalent to explicitation.

The concept of a precisiated natural language and the associated methodologies of computing with words and the computational theory of perceptions open the door to a wide-ranging generalization and restructuring of existing theories, especially in the realms of information processing, decision and control. In this perspective, what is very likely is that in coming years a number of basic concepts and techniques drawn from linguistics will  be playing a much more important role in scientific theories than they do today.
 

Dr. Zadeh holds honorary doctorates from Paul-Sabatier University, Toulouse, France; State University of New York, Binghamton, NY; University of Dortmund, Dortmund, Germany; University of Oviedo, Oviedo, Spain; University of Granada, Granada, Spain; Lakehead University, Canada; University of Louisville, KY; Baku State University, Azerbaijan; the Silesian Technical University, Gliwice, Poland; the University of Toronto, Toronto, Canada; the University of Ostrava, Ostrava, the Czech Republic; the University of Central Florida, Orlando, FL; the University of Hamburg, Hamburg, Germany; and the University of Paris(6), Paris, France.   Dr. Zadeh has single-authored over  two hundred papers and serves on the editorial boards of over fifty journals. He is a member of the Advisory Board, Fuzzy Initiative, North Rhine-Westfalia, Germany; Advisory Board, Fuzzy Logic Research Center, Texas A&M University, College Station, Texas; Advisory Committee, Center for Education and Research in Fuzzy Systems and Artificial Intelligence, Iasi, Romania; Senior Advisory Board, International Institute for General Systems Studies; the Board of Governors, International Neural Networks Society; and is the Honorary President of the Biomedical Fuzzy Systems Association of Japan and the Spanish Association for Fuzzy Logic and Technologies. In addition, he is a member of the International Steering Committee, Hebrew University School of Engineering; a member of the Advisory Board of the National Institute of Informatics, Tokyo; a member of the Governing Board, Knowledge Systems Institute, Skokie, IL; and an honorary member of the Academic Council of NAISO-IAAC.

Top

Evolutionary Optimization and Constraint Handling

Xin Yao
Chair Professor, Computer Science
University of Birmingham, United Kingdom
E-mail: X.Yao@cs.bham.ac.uk 

Evolutionary algorithms (EAs) have been used widely in optimization with success. However, it is not always clear why and how an EA works. This talk gives two case studies of how a deeper understanding of evolutionary computation techniques can further our understanding of EA's behaviors and guide us in developing new techniques. The first study analyses the impact of search operators, e.g., mutation, on EA's search behaviors. The importance of search step size is emphasized. The second study examines the penalty function approach to constraint handling. It is shown how a simple and effective constraint handling technique can be developed based on a better understanding of how the penalty function approach works in constrained optimization.

Biography:

Xin Yao's major research interests include evolutionary computation (especially evolutionary learning and optimization), neural networks ensembles, stochastic search algorithms (e.g., simulated annealing), evolvable hardware, and computational time complexity of evolutionary algorithms.

Top