Data processing: artificial intelligence – Neural network – Learning method
Reexamination Certificate
1999-02-02
2004-09-14
Davis, George B. (Department: 2121)
Data processing: artificial intelligence
Neural network
Learning method
C706S039000, C709S203000
Reexamination Certificate
active
06792412
ABSTRACT:
FIELD OF THE INVENTION
This invention relates to a system and method for controlling information output based on user feedback regarding that information. More particularly, the invention relates to a computer network-based neural network system that controls information provided as output from the system based on learned experience gained from user feedback regarding the value of the information.
BACKGROUND OF THE INVENTION
With the proliferation of network connectivity between users of computer systems, a vast amount of information has become readily available to those users. Indeed, many call recent times the “information age” because of the ease of gaining access to information from homes and businesses throughout the world. As the amount of available information increases, users have become overwhelmed because they simply do not have enough time or sufficient resources to absorb all of that information in any meaningful way.
Accordingly, users have created information filters of various types in an attempt to filter through the large value of information to derive useful information. Some electronic mail receiving programs parse incoming messages and filter out messages based on content or the initiator of the message. Search engines have been created in an attempt to identify relevant web pages on the Internet out of the millions of such pages that exist and the vast numbers of new pages that are created on a daily basis.
These systems all suffer from various drawbacks. To filter based on some characteristic of an electronic mail message, the electronic mail filtering system requires that such attributes be determined beforehand and provided for the filtering. The filtering system then filters every message having designated attributes. That requires identification of those attributes that are to be excluded before being able to create the filter. It is often difficult to determine the proper attributes to be used as the basis for the filtering.
Similarly, in some existing chat rooms, a user may select participants from whom to receive information. Other persons are excluded, even if the information they might provide would be useful to that user and even if the other participants in the chat room find that information to be useful. Such selections are inflexible to changing circumstances and again, like the electronic mail example, require the user to select those participants to include or exclude from the chat room without regard to the substance of the contribution those participants will make.
In other situations, web pages may include a number of links to other web pages. Sometimes these pages may contain hundreds or thousands of links and often new links being added. It may be difficult for a user to find or remember the links that were previously useful the next time the user opens that link directory page with such a large volume of links.
Further, although search engines exist to enable users to locate pages based on key words contained in the page, these engines also suffer from drawbacks. Many pages incorporate a large volume of words so that those pages will provide “hits” to the search engines that include those words. Often the page may contain such words even though the subject matter of the page may have little or nothing to do with the words that are included. Inclusion of words in this manner may render current search engines less useful in identifying pages with information that is likely to be useful to the user.
Other drawbacks also exist with current systems.
SUMMARY OF THE INVENTION
Accordingly, it is an object of the present invention to overcome these and other drawbacks of current systems.
One object of the present invention is to provide a system and method for delivering information over a network which selects information for delivery based on indications of what subject matter a recipient is likely to find useful.
Another object of the present invention is to provide a system and method for delivering information to users over a network based on learned preferences of users of that information.
Another object of the present invention is to provide a system and method for developing affiliations between information sources, information recipients, and information content.
Another object of the present invention is to provide a system and method for employing feedback about information provided over a network which is used to determine what information to provide to particular users in the future.
Another object of the present invention is to provide a system and method for automatically determining preferences of recipients and users of information distributed over a network.
Another object of the present invention is to provide a system and method for automatically determining relationships between attributes of information and user determinations regarding the usefulness of that information.
Accordingly, a system and method is provided for accomplishing these and other objects of the present invention. According to the present invention, a neural network system may be provided between a plurality of information sources and a plurality of information recipients. The neural network learns what information to provide from the information sources to the information recipients based on feedback provided by the feedback recipients regarding the usefulness of the information.
The information sources may comprise electronic mail providers, chat participants, and/or web page links, for example. The recipients may comprise electronic mail recipients, chat participants, and/or web page viewers, for example. Also, the input upon which the neural network bases its determination, for example, may comprise attributes of the electronic mail providers, chat participants, key words, and/or web pages. Accordingly, the present invention is applicable to a wide range of network-based applications and it enables the development of information sharing communities based on attributes and preferences of the recipients and providers. The system may also be hierarchical and connected to other learning systems to provide a scalable information distribution network.
Accordingly, the system may comprise a vast interconnection of information sources and recipients including news services, electronic mail, chat and web-based content. The system may bring together people and enhance the relevance of the information that reaches users on the system. The system may provide relevant electronic mail messages, help users find appropriate web pages through search engines and link directory pages, and may provide improved chat discussions by reducing the amount of non-useful content.
Although some objects and advantages of the present invention have been described, other advantages of the present invention will be apparent from reviewing the detailed description and drawings of the present invention set forth herein.
REFERENCES:
patent: 4999833 (1991-03-01), Lee
patent: 5390281 (1995-02-01), Luciw et al.
patent: 5410344 (1995-04-01), Graves et al.
patent: 5446891 (1995-08-01), Kaplan et al.
patent: 5504675 (1996-04-01), Cragun et al.
patent: 5515098 (1996-05-01), Carles
patent: 5636346 (1997-06-01), Saxe
patent: 5727129 (1998-03-01), Barrett et al.
patent: 5754938 (1998-05-01), Herz et al.
patent: 5754939 (1998-05-01), Herz et al.
patent: 5758257 (1998-05-01), Herz et al.
patent: 5765028 (1998-06-01), Gladden
patent: 5781698 (1998-07-01), Teller et al.
patent: 5781703 (1998-07-01), Desai et al.
patent: 5790935 (1998-08-01), Payton
patent: 5802253 (1998-09-01), Gross et al.
patent: 5822744 (1998-10-01), Kesel
patent: 5835087 (1998-11-01), Herz et al.
patent: 5991735 (1999-11-01), Gerace
patent: 6085178 (2000-07-01), Bigus et al.
patent: 6154783 (2000-11-01), Gilmour et al.
patent: 6266664 (2001-07-01), Russell-Falla et al.
patent: 6321205 (2001-11-01), Eder
El-Zoghabi et al, “AClient-Server Computing Model for Heterogeneous Distributed Management”, IEEE, Apr. 1993.*
The Delta Rule, http://uhavax.hartford.edu, printed on Jan. 4, 1999 (one page).
Autonomy Agentware Technology Wh
Pope Ivan
Sullivan Alan
Davis George B.
Moazzam & Latimer LLP
LandOfFree
Neural network system and method for controlling information... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Neural network system and method for controlling information..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Neural network system and method for controlling information... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3244404