Data processing: database and file management or data structures – Database design – Data structure types
Reexamination Certificate
2000-10-27
2004-11-09
Alam, Shahid (Department: 2172)
Data processing: database and file management or data structures
Database design
Data structure types
C707S793000, C707S793000, C707S793000, C704S270100, C386S349000
Reexamination Certificate
active
06816858
ABSTRACT:
CROSS-REFERENCE TO A RELATED PATENT APPLICATION
This patent application is related to commonly-assigned U.S. patent application Ser. No. 09/627,555, filed Jul. 28, 2000, to Bolle et al., entitled “Apparatus, System and Method for Augmenting Video Information Streams with Relevant Information”, the disclosure of which is incorporated by reference herein in its entirety.
FIELD OF THE INVENTION
This invention relates generally knowledge management methods and apparatus and, more specifically, the invention relates to knowledge management of information streams to determine knowledge concepts present in a content of an information stream and to determine additional or collateral information that is related to the content of the information stream.
BACKGROUND OF THE INVENTION
An information stream is a source of information where the information has a time-based component, and where the information “flows” from a source to a destination. The most common example of an information stream is spoken discourse (i.e., speech). The speaker is the information source, the listener is the destination, the content of the speech (the actual words) contains or represents the information, and the audible sound pressure wave produced by the speaker's mouth transmits the information from the speaker to the listener. The sound wave travels over time and must be processed in real-time (i.e., heard) by the listener. If the listener does not process the sound wave as it is received, the speech will be lost and the listener will not receive the information.
Other kinds of information streams include, for example, television broadcasts, telephone conversations, and computer network-based communications. An important feature of an information stream is that the information is transmitted over time and must be processed in real-time as it is received. Of course, this processing may include capture of the information (e.g., into a computer file) for further processing off-line at a later date.
Information streams are a valuable resource in the practice of knowledge management. Knowledge management is an activity that includes processes and technologies for capturing intellectual capital and making it easily accessible for reuse and exploitation (see, for example, Davenport and Prusak, “Working Knowledge”, Harvard Business School Press, Boston, 1998).
Many knowledge management tools exist that operate on textual information, or documents. The most basic operation is to index and search the documents using a text retrieval system (see, for example, Baeza-Yates and Ribeiro-Neto, “Modem Information Retrieval”, ACM Press, New York, 1999). More advanced operations on documents include automatic clustering, automatic classification, and automatic extraction of concepts and named entities from documents. One product that provides tools to perform all of these tasks on a collection of documents is the IBM Intelligent Miner for Text (see U.S. Pat. No. 5,832,480).
All of these previously described document processing tasks may be further refined with user profiles. A user profile describes a particular interest or set of interests on behalf of the user. The profile is used to filter or modify the various document processing tasks so that the results more closely match the interests of the end user.
The convergence of information streams and knowledge management occurs naturally in two important contexts: meetings and data broadcasting. Meetings have a variety of incarnations, with the most common being a face-to-face meeting between two or more individuals. The meeting will minimally include a spoken discourse information stream, and may additionally include other documents, such as an agenda, a visual presentation, and notes (i.e., meeting minutes). Other incarnations of meetings include sales presentations, teleconferences, video conferences, email exchanges, chat sessions, and help desk call sessions. For prior art related to meetings, see U.S. Pat. Nos. 5,890,131, 5,786,814, 6,018,346 and 5,465,370.
Data broadcasting is the process of encoding data in a television broadcast signal (in addition to the traditional video and audio signals). Both analog and newer digital television channels have unused bandwidth that can be used to transmit arbitrary data. This data may or may not be related to the accompanying audio/video broadcast. With the incorporation of data broadcasting, a television broadcast signal becomes a very rich information stream comprising audio, video, and data. For prior art related to data broadcasting, see U.S. Pat. Nos. 5,887,062 and 6,031,578.
The emergence of the World Wide Web (WWW or simply Web) as an information and entertainment media is generating many changes in the more traditional media of broadcast television. In particular, broadcasters have begun to link these two media together to create a much richer television viewing experience. For example, television programs may display URLs that point to Web sites related to the program. A next phase of linkage will enable set top boxes and TV tuner computer cards to become more prevalent. Such devices will allow broadcasters to send Web content with the television broadcast and display the audio/video program in an integrated fashion with the Web content.
This tighter integration of broadcast television and the Web presents a number of challenges, with one of the more difficult challenges being how to identify the information that should be broadcast with the television program. Currently, program producers manually identify the information to be broadcast. This process may be supported by software that aids in scheduling the data broadcast, or software that automatically accesses databases to obtain, for example, stock quotes. Nevertheless, the overall information seeking and selection process is manual.
This approach has several disadvantages. First, it is slow and expensive. Second, there is no mechanism to tie additional information into a live broadcast, where the time at which a particular topic is discussed is not known beforehand. Currently, if a significant event (e.g., a natural disaster occurs during a broadcast of the daily news), the producers have a difficult time just reporting the event, and in general may have no time to find background information. Third, with the advent of set top boxes, users may wish to customize the information displayed on their TV set. For example, one person may wish to see only sports-related information, while another may wish to choose news that is related to a specific geographic location.
One problem of particular interest to the teachings of this invention is most closely related to efforts related to Topic Detection and Tracking (TDT). Reference in this regard can be had to J. Allan, J. Carbonell, G. Doddington, J. Yamron, and Y. Yang, “Topic Detection and Tracking Pilot Study: Final Report”.
Proceedings of the DARPA Broadcast News Transcription and Understanding Workshop
, pp. 194-218. In TDT, the goal is to analyze news broadcasts (text articles or text transcripts generated automatically from audio and video) and to identify previously unseen news events, or topics. Topics are then tracked by identifying subsequent news stories covering the same event. This is accomplished using a variety of off-line text processing, language modeling, and machine learning algorithms. However, TDT is not a real-time system, so it cannot annotate a live broadcast with collateral information, and furthermore is basically limited to topic detection.
As was stated above, one information retrieval and text analysis technique includes the IBM Intelligent Miner for Text, “www-4.ibm.com/software/data/iminer/fortext/”. Reference may also be had to C. D. Manning and H. Schutze, “Foundations of Statistical Natural Language Processing”,
MIT Press
, 1999. However, neither of these approaches is specifically adapted to support on-line processing of streaming text data.
A number of commercial systems exist that support the manual addition of data to a broadcast signal (see, for example, Wave Systems Corporation a
Brown Eric William
Coden Anni Rosa
Alam Shahid
Ehichioya Fred
International Business Machines - Corporation
Ohlandt Greeley Ruggiero & Perle L.L.P.
Percello Louis J.
LandOfFree
System, method and apparatus providing collateral... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with System, method and apparatus providing collateral..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and System, method and apparatus providing collateral... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3333216