Multi-agent collaborative architecture for problem solving...

Education and demonstration – Question or problem eliciting response – Response of plural examinees communicated to monitor or...

Reexamination Certificate

Rate now

  [ 0.00 ] – not rated yet Voters 0   Comments 0

Details

C434S118000, C434S323000, C434S362000

Reexamination Certificate

active

06690914

ABSTRACT:

BACKGROUND OF THE INVENTION
The present invention relates generally to systems for tutoring and problem solving and, more particularly, to a multi-agent collaborative architecture for a computer system for tutoring and problem solving.
Computer systems that provide scenario-based training for individuals are known. Such a system is illustrated by U.S. Pat. No. 5,597,312 (BLOOM et al.). In addition, systems that provide computer-assisted tutoring in the medical field are also known. Such a system is disclosed in U.S. Pat. No. 4,945,476 (BODICK et al.). These systems, however, fail to take advantage of the benefits of “blackboard” computer system architectures.
The word “blackboard,” as used with reference to computer system architectures, came to be borrowed from the medical trauma domain where for a long time trauma doctors have diagnosed and treated patients coming into the hospital. Patients who come into a trauma ward often have serious injuries or illnesses and require immediate attention. In such instances, it is impossible for a single physician to handle the entire task of treating the patient. This is because each physician is an expert in a particular area and that no one single area is enough to diagnose and treat the patient. So, there must be a consorted effort on the part of a team of doctors to handle patients. If the doctors were all to grab at the patient and start their diagnosis/findings or speak at the same time while communicating their diagnosis/findings the result would be chaos (or the loss of information possibly leading to a misdiagnosis).
The system invented to keep order was to have a white note pad that was attached to the stretcher of the patient on which each doctor wrote his diagnosis/findings and what tests and medication was recommended so that each doctor could see what all the other previous doctors had said and prescribed. This was extremely important because of the conflicts in treatment and diagnosis. Further, it was very important to maintain an order in which the patent was reviewed (i.e.: the brain specialist would need blood pressure information etc. before he could pass a diagnosis or post his findings). This general scheme of distributed co-operative problem solving came to be adopted by the field of artificial intelligence (AI) under the broad term of blackboard architectures.
Computer systems that employ blackboard architectures to provide tutoring and diagnosis/problem solving exist in the prior art. For example, U.S. Pat. No. 5,448,722 (LYNNE et al.) shows that computer-based diagnostic systems using both multiple electronic blackboards and diagnostic modules are known. Indeed, U.S. Pat. No. 5,101,362 (SIMOUDIS) states that typical expert systems include a knowledge base, a working memory, which is also termed a “blackboard”, and a control portion or inference engine. The Simoudis '362 patent illustrates an expert system with multiple knowledge sources effectively performing operations on data present in the blackboard.
The patents cited and described above, however, fail to disclose an architecture that employs multiple agents and blackboards in a tutorial application. A need thus exists for an architecture that can provide the advantages and benefits of multi-agent collaborative processing in a tutorial environment. More specifically, a need exists for an architecture for a tutoring system that employs multiple agents and blackboards and that includes a tutoring agent, a student database, a training/testing database and multiple domain dependent agents, with each of the domain dependent agents having expertise in a specific field (domain).
In addition, prior art tutoring systems tend to be information-centric as opposed to user-centric. User-centric systems may be tuned to cater more to a specific user's needs and thus provide a user with a greater retention of information. A need thus exists for an architecture that may enable a tutoring system to be user-centric.
Many prior art systems are also difficult to modify and update. Often the complete system software must be downloaded and/or replaced. A system with a modular architecture featuring numerous agents permits ease of updating and modification in that specific agents may be modified or updated independently or entire agents may be added or eliminated altogether. In addition, a modular architecture facilitates role-playing in that users may be easily substituted for agents of the system.
Many prior art systems also do not employ transmission control protocol/internet protocol (TCP/IP) links to permit the architecture agents to communicate with one another. Such an arrangement provides the capability for geographically diverse users to access the system simultaneously for role-playing or other interaction.
Accordingly, it is an object of the present invention to provide an architecture for problem solving and tutoring that features multiple agents and blackboards.
It is another object of the present invention to provide an architecture for problem solving and tutoring that is user-centric.
It is another object of the present invention to provide an architecture for problem solving that is modular.
It is still another object of the present invention to provide an architecture that uses TCP/IP links to provide communication between its agents.
SUMMARY OF THE INVENTION
The present invention is an architecture for a system for tutoring and problem solving that includes a number of domain dependent agents, each including a knowledge base, a number of domain independent agents and a number of blackboards. One of each of the blackboards is assigned or dedicated to one of each of the domain dependent agents or domain independent agents. A blackboard agent transfers messages between the blackboards so that the domain dependent and domain independent agents may communicate with one another by posting messages to and reading messages from their dedicated blackboards. The domain dependent agents and domain independent agents communicate with their dedicated blackboards through transmission control protocol/internet protocol links
During a session, a token is generated and the domain independent agents include a control agent that selectively passes the token between the domain dependent and domain independent agents so that each functions only when it has the token. The domain independent agents also include a tutoring agent that communicates with the domain dependent and domain independent agents through the blackboards to enable a user of the computer system to select between the domain dependent agents for a tutoring session.
The architecture also includes a student database and a training/testing data repository or database and the plurality of domain independent agents include a database agent. The database agent provides data to the domain dependent and domain independent agents from the databases through the blackboards and receives data from the domain dependent and domain independent agents for posting on the student database. As a result, the actions of users during tutoring sessions may be recorded on the student database for tracking and future use.
The architecture also includes a user display and a graphical front-end, and the independent agents also include a user interface agent, for displaying information to the user, including data from the domain dependent agent knowledge bases and the training/testing data repository. In addition, the user interface agent also receives inputs from a user and directs them to appropriate domain dependent and domain independent agents through the blackboards.
The tutoring system may be installed on more than one computer which enables geographically diverse users to simultaneously role-play during a tutoring session.
The following detailed description of embodiments of the invention, taken in conjunction with the accompanying drawings, provide a more complete understanding of the nature and scope of the invention.


REFERENCES:
patent: 4370983 (1983-02-01), Lichtenstein
patent: 4464172 (1984-08-01), Lichtenstein
patent: 473

LandOfFree

Say what you really think

Search LandOfFree.com for the USA inventors and patents. Rate them and share your experience with other people.

Rating

Multi-agent collaborative architecture for problem solving... does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Multi-agent collaborative architecture for problem solving..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Multi-agent collaborative architecture for problem solving... will most certainly appreciate the feedback.

Rate now

     

Profile ID: LFUS-PAI-O-3327391

  Search
All data on this website is collected from public sources. Our data reflects the most accurate information available at the time of publication.