Intelligent & Adaptive Tools
Intelligent and adaptive (I&A) tools are computing programs, components, and systems that can reason about complex situations and adapt that reasoning to new information and unforeseen scenarios. Such intelligent and adaptive capabilities form the basis of effective decision-support software.
To provide such support, I&A tools require the following:
For I&A tools to execute sophisticated reasoning, they need expressive representations of the concepts, entities, and phenomena comprising their operational domain. These expressive representations—or models—go beyond traditional data-centric renditions and provide I&A tools with the contextual information needed to answer complex questions posed by end-users.
At the same time, these models are typically designed to accommodate new concepts and additional principles as such information emerges. In this way, then, information-rich models sustain the adaptive qualities of I&A tools.
Traditionally, software tools categorized as intelligent were engineered for specific scenarios, making their ultimate utility a matter of how closely their design aligned with the characteristics of a particular scenario. Such rigidity proved to be more problematic than productive, as these tools struggled to accommodate even the slightest variation in the scenarios for which they were designed.
I&A tools, on the other hand, are engineered with adaptation in mind. This adaptability derives from the structure of the information-rich models these tools reference as well as these tools' ability to decompose complex processes into manageable components, without sacrificing the contextual relationships that give these components meaning. Working on components of a process rather than the entire process, I&A tools can perform tasks that are not tightly bound to the information currently governing the total process. The result is a problem solving tool that is far more agile than traditional applications in its ability to adjust to the variable nature of real world problems.
Human/Computer Partnership (Collaboration)
In the past, intelligent tools were engineered to fully automate business activities without exposing either the logic (i.e., considerations, assumptions, constraints, etc.) or the subsequent decisions made to end-users. While this approach certainly relieved users from having to deal with the complexity and time-consuming aspects of particular tasks, it also left the analysis and decision-making to a computing facility that often lacked the expertise, experience, and knowledge of the world necessary to make effective decisions.
Adaptability techniques, including goal and task decomposition.
Intelligent and learning capabilities needed for communities of I&A tools to achieve their objectives. Emphasis on problem solving (i.e., logic, search, and heuristics), planning, communication, and negotiation techniques.
Semantically oriented models (or ontologies). Emphasis on concepts, principles, and techniques that enable representations to be expressive (i.e., sufficiently detailed to capture contextual information) yet adaptable (i.e., sufficiently abstract to evolve as information evolves).
A logistician uses intelligent agent technology to help her determine the best method for shipping materials.
In contrast, intelligent and adaptive tools are expressly designed to form partnerships with their human user(s), playing the role of team member rather than automated decision maker. In this role, these tools collaborate with human end-users to acquire additional or tacit information about their domain, provide human end-users with insight into the reasoning they apply, and involve human end-users as final arbitrators, whenever confronted with internally irresolvable conflicts. With such collaborative capabilities, I&A tools better assist human decision makers in dealing with complex problems.