Howard W. Beck

Classification as a query processing technique in the CANDIDE semantic data model

The use of classification and subsumption to process database queries is discussed. The data model, called CANDIDE, is essentially an extended version of the FL-1, KANDOR and BACK, frame-based knowledge representation languages. A novel feature of the approach is that the data-description language and data-manipulation language are identical, thus providing uniform treatment of data objects, query objects and view objects. The classification algorithm find the correct placement for a query object in a given object taxonomy. Tractability issues are explored, and the expressiveness of queries is compared with relational algebra. This data model has been implemented in POPLOG as the basis for a knowledge-base management system that includes an integrated natural-language query system.<<ETX>>

Knowledge mining by imprecise querying: a classification-based approach

Knowledge mining is the process of discovering knowledge that is hitherto unknown. An approach to knowledge mining by imprecise querying that utilizes conceptual clustering techniques is presented. The query processor has both a deductive and an inductive component. The deductive component finds precise matches in the traditional sense, and the inductive component identifies ways in which imprecise matches may be considered similar. Ranking on similarity is done by using the database taxonomy, by which similar instances become members of the same class. Relative similarity is determined by depth in the taxonomy. The conceptual clustering algorithm, its use in query processing, and an example are presented.<<ETX>>

Integrating natural language, query processing, and semantic data models

The design and implementation of a natural language query processing system is described. The system is based on the CANDIDE object-oriented data model, which consists of a collection of instances and classes. Classes are arranged in a generalization hierarchy. Each instance belongs to one or more classes. CANDIDE includes a terminological reasoning capability, known as the Classifier, which is used to process queries. Natural language queries are converted into CANDIDE objects and then processed by the Classifier. Issues in natural language processing are discussed, including the lexical acquisition problem. The approach is unique in terms of aiding natural language processing with a semantic data model.<<ETX>>

The National Plant Diagnostic Network

Following September 11, Americas attention and resources were refocused on homeland security. While emphasizing the security of facilities, such as airports, tourist attraction sites, and major public buildings, etc. Congress also recognized the vulnerability of its agricultural systems. On June 12, 2002, the President signed into law the Agricultural Bio-terrorism Protection Act of 2002. The Act covers both animal and plant production and directed the Secretary of Agriculture to develop a network linking plant and animal disease diagnostic facilities across the country. The National Plant diagnostic Network (NPDN) focuses on the plant diseases and pest aspects of the program. Its mission is to enhance national agricultural security by quickly detecting introduced pests and pathogens. This will be achieved by creating a functional nationwide network of public agricultural institutions with a cohesive, distributed system to quickly detect deliberately introduced, high consequence, biological pests and pathogens into our agricultural and natural ecosystems by providing means for quick identifications and establishing protocols for immediate reporting to appropriate responders and decision makers. The network is comprised of Land Grant University plant disease and pest diagnostic facilities across the United States. Lead universities have been selected and designated as regional Centers to represent five regions across the country. These Centers are located at Cornell University, Michigan State University, Kansas State University, University of Florida at Gainesville, and University of California at Davis. The National Agricultural Pest Information System (NAPIS), located at Purdue University, has been designated as the central repository for archiving select data collected from the regions. NAPIS maintains information from the Cooperative Agricultural Pest Survey (CAPS), a network of state agricultural organizations and universities that survey for invasive species. As a part of the NPDN, NAPIS will expand its collection of data on plant diseases and other pests. The system will provide a national perspective on agricultural pests through dynamic maps and reports of plant pest distribution. Currently the pest information system houses 1.3 million records on more than 3,800 organisms, and that number will grow significantly as the plant diagnostic network centers start feeding information into the national database in the Spring, 2004. The establishment of the network will provide the means necessary for ensuring all participating Land Grant University diagnostic facilities are alerted of possible outbreaks and/or introductions and are technologically equipped to rapidly detect and identify pests and pathogens. This will be accomplished by establishing an effective communication network between regional expertise, developing harmonized reporting protocols with the national diagnostic network participants, and cataloging pest and disease occurrence to be included in the national database. The NPDN national database at Purdue University will provide summary reports, distribution maps, pattern analysis, and data sets for use in other studies.

A GIS-based database management application for agroforestry planning and tree selection.

Agroforestry (the deliberate growing of trees or shrubs in rural lands) is being promoted in the United States as an alternative resource management system that can bring landowners economic benefits and provide environmental services such as reduced soil erosion, improved water quality and wildlife habitat. Landowners, farmers and extension agents need to be better informed about different agroforestry opportunities and potential tree species. The Florida Agroforestry Decision Support System (FADSS) was designed to aid in the dissemination of such information. FADSS utilizes a geographical information system (GIS) enabling the user to select a location of interest which is linked to spatial data on climate and soils characteristics for the state of Florida. The application also incorporates a database of over 500 trees and 50 tree attributes, forming a relational database. The application structure consists primarily of building database queries using Standard Query Language (SQL). SQL queries are constructed during run-time based on spatial parameters of a selected location, the type of agroforestry system desired, and production and management criteria provided by the user. Experts were interviewed to help develop queries used to select trees and other agroforestry species. Being a prototype, the application is built with a modular and flexible framework in which spatial data of different scales and/or regions as well as plant data may be easily incorporated. Among the major limitations encountered during the development of FADSS with major implications on future agroforestry decision support systems was the current lack of tree information relevant to agroforestry and the lack of research involving the assessment of suitable trees and their characteristics.

Incorporating natural language descriptions into modeling and simulation

We explore an approach to merging simulation and natural language in which conceptual structures are used which can repre sent the structure and meaning of sentences as well as basic mathematical concepts. Sentences can be transformed to these structures, and, via language generation, the structures can be transformed into mathematical equations. The process is illustrated using a text description of a plankton respiration model. The ap proach shows how qualitative natural language statements can be merged with both qualitative and quantitative modeling and simulation.

A conceptual clustering algorithm for database schema design

Conceptual clustering techniques based on current theories of categorization provide a way to design database schemas that more accurately represent classes. An approach is presented in which classes are treated as complex clusters of concepts rather than as simple predicates. An important service provided by the database is determining whether a particular instance is a member of a class. A conceptual clustering algorithm based on theories of categorization aids in building classes by grouping related instances and developing class descriptions. The resulting database schema addresses a number of properties of categories, including default values and prototypes, analogical reasoning, exception handling, and family resemblance. Class cohesion results from trying to resolve conflicts between building generalized class descriptions and accommodating members of the class that deviate from these descriptions. This is achieved by combining techniques from machine learning, specifically explanation-based learning and case-based reasoning. A subsumption function is used to compare two class descriptions. A realization function is used to determine whether an instance meets an existing class description. A new function, INTERSECT, is introduced to compare the similarity of two instances. INTERSECT is used in defining an exception condition. Exception handling results in schema modification. This approach is applied to the database problems of schema integration, schema generation, query processing, and view creation. >

Agricultural enterprise information management using object databases, Java, and CORBA

A software architecture based on an object database management system (ODBMS), Java, and Common Object Request Broker Architecture (CORBA) was applied to a variety of agricultural enterprise applications. The advantages and disadvantages of using object database are compared with conventional relational database management systems in complex applications. In a distributed object computing environment, a commercial ODBMS provides a repository for information on agriculture and natural resources in large organizations, Java provides a high-level programming environment, and CORBA provides a way of sharing applications and information across the Internet. Several applications are described, including an extension publication archive and an institutional accountability project. Results were that the object database offered significant advantages in modeling complex data applications, and the object design approach led to an overall increase in productivity by software developers. However, the technology in general suffers from a lack of maturity.

Interoperation of Organizational Data, Rules, Processes and Services for Achieving Inter-Organizational Coordination and Collaboration

Solutions to many complex problems that government organizations all over the world face today require these organizations to share, not only data and computing resources,but also policies, constraints, regulations, processes and services in order to achieve inter-organizational coordination and collaboration. This paper presents an integrated specification language and a user interface for collaborating government organizations to specify events of common interest, policies,constraints and regulations in the form of different types of knowledge rules, manual and automated services, and sharable workflow processes. A network system infrastructure for dynamic processing and interoperation of distributed rules and processes, and an event-triggered rule processing and process enactment technique are also described.

Natural language, cognitive models, and simulation

Models used in qualitative simulation are suitable for use as formal cognitive models, such as those involved in representing language meaning. In a series of examples, we explore the mapping between natural language expressions and formal models used in computer simulation. We present a theoretical representation of categories and word meanings in which cognitive models play an important role. The examples illustrate the use of a model in reasoning and discourse, the expression of temporal relationships, verbal descriptions of mathematical expressions, and the generation of qualitative descriptions of model behavior. This work can be applied in the process of software engineering as natural language specifications are transformed into models, or model results are interpreted and reported by natural language generators. Furthermore, models of various kinds are necessary in systems that use language.