Data processing: speech signal processing – linguistics – language – Speech signal processing – Recognition
Reexamination Certificate
1999-06-17
2001-02-20
Zele, Krista (Department: 2748)
Data processing: speech signal processing, linguistics, language
Speech signal processing
Recognition
Reexamination Certificate
active
06192338
ABSTRACT:
FIELD OF THE INVENTION
This invention relates generally to speech recognition, and more particularly to a natural language knowledge server for use as a speech recognition resource in a network.
BACKGROUND OF THE INVENTION
The Internet is a global network of interconnected networks and computers which enables users at computers coupled to the networks to exchange information including data, images, voice, recorded sound, and electronic mail. Computers connected to the Internet may be classified as client computers (“clients”) and server computers (“servers”). Servers are repositories of information. Internet users access the information present at the servers using client computers. Protocols establish the manner of information exchange between clients and other clients, between clients and servers, and between servers and other servers. The hypertext transfer protocol is one example.
Although natural language processing systems are beginning to emerge, capable of appropriate responses to spoken or typed queries, the integration of such capabilities into network applications is complex and time consuming. Customization for each application requires highly specialized knowledge outside the domain of normal application developers. Developing applications for performing speech recognition requires highly specialized knowledge.
In a conventional arrangement, a client computer is communicatively coupled to a single server computer according to a client-server architecture. Packets of data representing speech are passed from the client computer to the server computer. The server computer is capable of performing automatic speech recognition and also interpreting the results of the speech recognition. In the single “monolithic” server computer, the speech recognition function is tightly coupled to the function that interprets recognition results for a specific task domain. The two functions are likely accomplished in the same application running on the monolithic server computer.
While intimate knowledge about the task domain allows small improvements to be made in the speech recognition function, and knowledge of characteristics of the speech recognition function allows specific improvements in the result interpretation function, such improvements are made difficult by the complex interdependency of the application functions performed by the monolithic server computer. Shortcomings of this approach are that it is not easy to separate the two functions of speech recognition and result interpretation in the single monolithic server computer, and thus, the monolithic server computer cannot be applied to a new task without extensively redesigning both functions. For the same reason, as technology improves, it is often difficult to add new or improved functionality without a similar redesign of the single monolithic server computer.
SUMMARY OF THE INVENTION
An illustrative network resource system includes a first server which is capable of communicating with a client computer. The first server produces (i) a speech signal representing speech from a user at the client computer, and (ii) context information which indicates the kind of application the first server is running, the semantic context of the application, any prior knowledge relating to the user's request and a predefined format in which data are returned to the first server. A network knowledge server is in communication with and separated from the first server. The network knowledge server returns to the first server a text structure having one or more fields corresponding to such predefined format. The first server uses data from such one or more fields to determine a response to the user's speech.
Other features and advantages of the invention will become apparent from the following detailed description, taken in conjunction with the accompanying drawing, which illustrates, by way of example, the features of the invention.
REFERENCES:
patent: 5249293 (1993-09-01), Schreiber et al.
patent: 5475792 (1995-12-01), Stanford et al.
patent: 5513298 (1996-04-01), Stanford et al.
patent: 5615296 (1997-03-01), Stanford et al.
patent: 5632002 (1997-05-01), Hashimoto et al.
patent: 5664061 (1997-09-01), Andreshak et al.
patent: 5664126 (1997-09-01), Hirakawa et al.
patent: 5706507 (1998-01-01), Schloss
patent: 5790792 (1998-08-01), Dudgeon et al.
patent: 5867494 (1999-02-01), Krishnaswamy et al.
patent: 5884033 (1999-03-01), Duvall et al.
patent: 5918214 (1999-06-01), Perkowski
U.S. applications No. 08/833210, Dragosh et al., filed Apr. 14, 1997.
U.S. application No. 08/886136, Strahs filed Jun. 30, 1997.
Haszto Edward Dennis
Levin Esther
Marcus Stephen Michael
Pieraccini Roberto
AT&T Corp.
Opsasnick Michael N.
Zele Krista
LandOfFree
Natural language knowledge servers as network resources does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Natural language knowledge servers as network resources, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Natural language knowledge servers as network resources will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2571247