Web site and design © 2008 by G6G Consulting Group. All Rights Reserved. Most product content has been taken directly from manufacturer's web sites;
other product content is assembled by G6G Consulting Group. G6G welcomes any corrections and/or comments.
Product Feedback
* Required Field
*Your name:
*Email:
*Questions, comments, or feedback:
    Acquire

    Category  Intelligent Software>Expert (Knowledge Based)
    Systems/Tools

    Abstract  Acquire is a software package that helps ordinary people
    build expert (knowledge based) system applications. Acquire's
    structured methodology lets you build a knowledge base while the
    inference engine lets others use that knowledge. Programmers use
    the Acquire SDK (Software Development Kit) (see G6G Product
    Number 20041) to integrate knowledge bases with any software
    application and the Acquire Service (NT/2000/XP) (see G6G Product
    Number 20041) to operate them in a client/server environment.

    Acquire includes A knowledge base editor and integrated inference
    engine. Acquire Advantages/Capabilities include:

    1) A structured approach - Product uses a step-by-step methodology to
    acquire and represent knowledge eliminating the need for specialized
    training; 2) Knowledge acquisition by pattern recognition - People react
    to situations they recognize - this is the basis of knowledge in both
    people and Acquire. Rules specify the pattern of information that
    should be recognized in a situation and the reaction it should invoke.
    These actions may generate new patterns recognized by other rules,
    and so on. Thus, an expert system captures the pattern-based
    knowledge that is applied in a sequence of 'recognize-act' cycles; 3)
    Complete and consistent knowledge bases - The thoroughness of
    step-by-step development based on pattern-recognition minimizes
    errors because all situations that a user can recognize are specified
    and handled. Appropriate structuring of the knowledge-base
    eliminates irrelevant patterns from consideration; 4) Comparative
    handling of uncertainty - Many systems handle uncertainty with
    numerical confidence values but people do Not specify such absolute
    numbers reliably. Yet they are over-confident in them, and this leads to
    unreliable results in an application. Acquire handles uncertainty
    comparatively, consistent with our natural way of thinking, resulting in
    superior performance for applications.

    Acquire realizes these advantages/capabilities through its
    methodology and the steps it follows to help you build a knowledge-
    base.

    Acquire Methodology - Acquire is based upon two (2) fundamental
    design concepts:

    1) Expertise is normally applied quickly and unconsciously so it is
    likely based on recognizing and reacting to patterns of information
    (Pattern Recognition) rather than recalling and fitting rules to a
    situation.

    2) People are much better at making relative rather than absolute
    numerical judgments, so conflicts should be resolved by comparative
    (Relative Judgments) rather than quantitative methods.

    Pattern Recognition: The Basis of Expertise - Acquire provides an
    editor to build production rules, but it also provides the capability to
    represent expert behavior as singular, specific patterns of inputs
    matched to one consequence. Input pattern and consequence
    matches are enumerated in an action table.

    Relative Judgments: Contexts, Preferences and Biases - Acquire uses
    three (3) methods to resolve conflicts between rules - 1) Contexts
    avoid conflicts in the first place; 2) Preferences rank rules; 3) Biases
    favor conclusions. These qualitative solutions capitalize on our skill at
    making relative judgments.

    Knowledge Acquisition with Acquire - The structured knowledge
    acquisition methodology implemented in Acquire helps to maintain the
    perspective and focus of attention needed to build a thorough and
    consistent knowledge-based application. This structured approach
    involves the following steps:

    1) Build Objects - The first step is to name the objects (data, variables,
    attributes, concepts, hypotheses, conclusions, etc.) referred to in the
    subject area. For each object, assign a value-set that represents the
    range of meanings that the object can assume. For example, the
    object "Temperature" could be assigned the values freezing, cold,
    warm, or hot; or it could be a number alone; it could be a number
    mapped onto one of the symbolic values.

    2) Link Objects - The second step is to structure objects into a support
    network by linking those that influence each other. Which objects can
    affect (i.e., determine the value of) the current object? Which ones does
    it in turn affect? Linking all related objects produces a object graph that
    represents a general model of the domain without any cluttering detail
    (i.e., value assignments or rule details). The object network provides
    an overview of the way that knowledge is organized in the domain.

    3) Build Rules - The third step is to build rules from related objects.
    Select an object and choose from its supporting objects those that
    should be involved in determining its value, in this rule. Now, the utility
    of specifying links between objects is apparent: it allows you to work
    with manageable units in the knowledge-base. The relationship
    between these objects can be specified in either a production rule or
    an action (decision) table. This step results in graph of related rules.
    However, whereas the object network provides a general overview of
    the domain, the rule network provides the knowledge of how to actually
    retrieve the appropriate knowledge in a specific situation. This is the
    difference underlying a knowledge management system which links
    documents and an expert system which additionally knows how to
    apply the knowledge.

    4) Resolve Conflicts - Conflicts between rules occur when one or more
    rules try to set an object to different values on the same cycle. Conflicts
    should be resolved by preventing such situations in advance through
    appropriate rule conditions that enforce mutual exclusion between
    competing rules. If the expertise in the domain is Not sufficient to allow
    this, then rule preferences can specify the suitability of each rule within
    the set of rules actually involved and the best rule is selected. If this
    fails, then value biases (e.g., highest, average, most common, etc.)
    can determine which value should be assigned by the rule.

    5) Build User Interface - The final step in building the knowledge base
    is to build a user interface for entering information and displaying
    results.

    The above steps help you manage the knowledge acquisition process
    by showing where to start and how to progress using steps that are
    implicitly required, so that you can focus on the knowledge to be
    captured. Each step doesn't need to be completed before starting the
    next; you can work on small units of related objects. It may also be
    useful to complete some details (e.g., object values, rule
    specifications) after sketching out general relationships by linking
    objects and packaging them into rule.

    System Requirements  Windows 95, 98, NT, ME, 2000, or XP. Under
    special arrangements, Acquire can be made available for most Unix
    and Linux platforms.

    Manufacturer   Home office; see web site for international locations.

    Acquired Intelligence Inc.
    205 - 1095 McKenzie Avenue
    Victoria, B.C., Canada, V8P 2L5  
    Canada
    Telephone:  (250) 479-8646
    Fax:  (250) 479-0764
    Email: info@aiinc.ca
    sales@aiinc.ca
    webmaster@aiinc.ca
The G6G Directory of Omics and Intelligent Software
Search www.G6G-SoftwareDirectory.com