Voice processing system, method for allocating acoustic and/or written character strings to words or lexical entries

• US 8,498,859 B2
• Filed: 11/12/2003
• Issued: 07/30/2013
• Est. Priority Date: 11/13/2002
• Status: Active Grant

• Current Assignee: Bernd Schonebeck
• Original Assignee: Bernd Schonebeck
• Inventors: Schnebeck, Bernd
• Primary Examiner(s): Shah, Paras D
• Application Number:
  US10/534,658
• Publication Number:
  US 20060100851A1
• Time in Patent Office: 3,548 Days
• Field of Search: 704/231, 704/251, 704/257, 704/270, 704/275, 704/9, 704/254, 923/1
• US Class Current: 704/9
• CPC Class Codes: G06F 40/35 Discourse or dialogue representation
• CPC Class Codes: G10L 15/1815 Semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning
• CPC Class Codes: G10L 15/1822 Parsing for meaning understanding
• CPC Class Codes: G10L 15/22 Procedures used during a speech recognition process, e.g. man-machine dialogue
• CPC Class Codes: H04M 2201/40 using speech recognition speech recognition per se G10L15/00
• CPC Class Codes: H04M 2203/2061 Language aspects
• CPC Class Codes: H04M 3/493 Interactive information services, e.g. directory enquiries ; Arrangements therefor, e.g. interactive voice response [IVR] systems or voice portals
• CPC Class Codes: H04M 3/4936 Speech interaction details speech recognition per se G10L15/00
• CPC Class Codes: H04M 3/527 Centralised call answering arrangements not requiring operator intervention

Abstract

A language-processing system has an input for language in text or audio, as a message, an extractor operating to separate words and phrases from the input, to consult a knowledge base, and to assign a concept to individual ones of the words or phrases, and a connector operating to link the concepts to form a statement. In some cases there is a situation model updated as language is processed. The system may be used for controlling technical systems, such as robotic systems.

Claims

1. A system for automatically recognizing and further processing meaning of a linguistically provided information, comprising:

a processor executing routines providing;

an extractor which extracts a concept and a concept category from a knowledge base for each word contained in said linguistically provided information, each extracted concept corresponding to isolated meaning of one of the words; and

a connector which connects the extracted concepts to obtain a reconstructed meaning of said linguistically provided information;

wherein;

an artificial language intelligence coordinates an information exchange between the connector and the knowledge base, which contains a pool of cognitive routines;

the connector first transmits a request code to the artificial language intelligence, the request code containing at least information regarding the connector and the extracted concept categories to be processed;

the artificial language intelligence subsequently processes the request code, thereby selecting a cognitive connection routine from the pool of cognitive routines, said cognitive connection routine determined by the combination of information contained in the request code, extracts information from the knowledge base in consideration of the extracted concept categories to be processed, and transmits the cognitive connection routine and the information extracted from the knowledge base to the connector; and

the connector connects the extracted concepts in consideration of said cognitive connection routine and said information extracted from the knowledge base, following which a meaningfully reconstructed meaning of said linguistically provided information is conveyed to a conflict module, the conflict module compares the reconstructed meaning of said information with a current situation context contained in a situation module, which is part of the knowledge base, by using an appropriate comparison routine, by the following process;

the conflict module first transmits a request code to the artificial language intelligence, the request code containing at least information regarding the conflict module, the extracted concept categories to be processed and the situation module;

the artificial language intelligence subsequently processes the request code, thereby selecting a cognitive comparison routine from the pool of cognitive routines, said cognitive comparison routine determined by the combination of information contained in the request code, extracts information from the situation module in consideration of the extracted concept categories to be processed;

transmits the cognitive comparison routine and the information extracted from the situation module to the conflict module; and

the conflict module compares the reconstructed meaning of said linguistically provided information with said information extracted from the situation module in consideration of said cognitive comparison routine.

2. The system according to claim 1, comprising a feedback module, wherein concepts for which no linking is accomplished are conveyed to the feedback module, which brings them to the attention of a user of the system.

3. The system according to claim 1, comprising a feedback module, wherein concepts which are contained in said information extracted from the knowledge base and which are added to the linguistically provided information by the connector, are conveyed to the feedback module which brings them to the attention of a user of the system.

4. The system according to claim 2, wherein the feedback module initiates a query after receiving said unlinkable or added concepts.

5. The system according to claim 3, wherein the feedback module initiates a query after receiving said unlinkable or added concepts.

6. The system according to claim 2, wherein the linguistically provided information is evaluated as a meaningless statement if the quantity of unlinked concepts exceeds a predefined number.

7. The system according to claim 1 wherein unfamiliar concepts are deposited via a learning module in the knowledge base.

8. The system according to claim 1, comprising a feedback module, wherein a valid reconstructed meaning of said linguistically provided information that can nevertheless not be realized on the basis of the current situation is recognized by the conflict module and is transferred to the feedback module which brings it to the attention of a user of the system.

9. The system according to claim 1, comprising a feedback module, wherein a valid reconstructed meaning of said linguistically provided information that can nevertheless not be realized on the basis of the current situation is recognized by the conflict module and is transferred to the feedback module which brings it to the attention of a user of the system.

10. The system according to claim 1, wherein a meaning of said linguistically provided information, which is possible in the current situation context, is subjected to a risk analysis by an expertise module, wherein the expertise module compares the reconstructed meaning of said linguistically provided information with expert knowledge contained in a expert knowledge module, which is part of the knowledge base, by using an appropriate comparison routine.

11. The system according to claim 1, wherein a meaning of said linguistically provided information, which is possible in the current situation context, is subjected to a risk analysis by an expertise module, wherein the expertise module compares the reconstructed meaning of said linguistically provided information with expert knowledge contained in a expert knowledge module, which is part of the knowledge base, by using an appropriate comparison routine.

12. The system according to claim 10, wherein:

the expertise module first transmits a request code to the artificial language intelligence, the request code containing at least information regarding the expertise module, the extracted concept categories to be processed and the expert knowledge module;

the artificial language intelligence subsequently processes the request code, thereby selecting a cognitive comparison routine from the pool of cognitive routines, said cognitive comparison routine determined by the combination of information contained in the request code, extracts information from the expert knowledge module in consideration of the extracted concept categories to be processed, and transmits the cognitive comparison routine and the information extracted from the expert knowledge module to the expertise module; and

the expertise module compares the reconstructed meaning of said linguistically provided information with said information extracted from the expert knowledge module in consideration of said cognitive comparison routine.

13. The system according to claim 11, wherein:

the expertise module first transmits a request code to the artificial language intelligence, the request code containing at least information regarding the expertise module, the extracted concept categories to be processed and the expert knowledge module;

the artificial language intelligence subsequently processes the request code, thereby selecting a cognitive comparison routine from the pool of cognitive routines, said cognitive comparison routine determined by the combination of information contained in the request code, extracts information from the expert knowledge module in consideration of the extracted concept categories to be processed, and transmits the cognitive comparison routine and the information extracted from the expert knowledge module to the expertise module; and

the expertise module compares the reconstructed meaning of said linguistically provided information with said information extracted from the expert knowledge module in consideration of said cognitive comparison routine.

14. The system according to claim 12, comprising a feedback module, wherein a valid reconstructed meaning of said linguistically provided information that can nevertheless not be realized on the basis of the information extracted from the expert knowledge module are recognized by the expertise module and are transferred to the feedback module which brings them to the attention of a user of the system.

15. The system according to claim 13, wherein a valid reconstructed meaning of said linguistically provided information is processed via a virtual realization module.

16. The system according to claim 15, the virtual realization module comprising a modification module which establishes any changes that are connected with the reconstructed meaning of said linguistically provided information and transmits them to the situation module for updating the latter in an ongoing manner.

17. The system according to claim 15, the virtual realization module comprising an anticipation module, wherein the anticipation module anticipates consequences of actions and events contained in the reconstructed meaning of said linguistically provided information in consideration of world knowledge contained in a world knowledge module, which is part of the knowledge base, by using an appropriate anticipation routine.

18. The system according to claim 17, wherein;

the anticipation module first transmits a request code to the artificial language intelligence, the request code containing at least information regarding the anticipation module, the extracted concept categories to be processed and the world knowledge module;

the artificial language intelligence subsequently processes the request code, thereby selecting a cognitive anticipation routine from the pool of cognitive routines, said cognitive anticipation routine is determined by the combination of information contained in the request code, extracts potential links to subsequent events for the reconstructed meaning of said linguistically provided information from the world knowledge module, and transmits the cognitive anticipation routine and, if the search is successful, the located links to the anticipation module; and

the anticipation module connects the reconstructed meaning of said linguistically provided information with said links extracted from the world knowledge module in the context-bound manner, by using said cognitive anticipation routine.

19. The system according to claim 17, comprising a feedback module, wherein anticipated subsequent events are conveyed to the feedback module, which brings them to the attention of a user of the system.

20. The system according to claim 18, wherein information made 5 available by the anticipation module is integrated in the situation model contained in the situation module assuring a current updating of the situation model on an ongoing basis.

21. The system according to claim 17, wherein the anticipated events are virtually realized by the virtual realization module, wherein:

the virtual realization module first transmits a request code to the artificial language intelligence, the request code containing at least information regarding the virtual realization module, the extracted concept categories to be processed and the world knowledge module;

the artificial language intelligence subsequently processes the request code, thereby selecting a cognitive virtual realization routine from the pool of cognitive routines, said cognitive virtual realization routine is determined by the combination of information contained in the request code, extracts information from the world knowledge module in consideration of the extracted concept categories to be processed, and transmits the cognitive virtual realization routine and the information extracted from the world knowledge module to the virtual realization module; and

the virtual realization module virtually realizes the reconstructed meaning of said linguistically provided information in consideration of said cognitive processing routine and said information extracted from the world knowledge module.

22. The system according to claim 15, comprising a command generation module, wherein the reconstructed meaning of said linguistically provided information, after undergoing the virtual realization, is converted by the command generation module into control commands for a downstream technical installation or robot.

23. The system according to claim 1, wherein the extractor reduces the linguistically provided information to basic forms.

24. The system according to claim 1, comprising special slots for implementing user-specific expert knowledge, wherein the meaning of a new linguistically provided information is first reconstructed, is then compared with the content of the existing knowledge base, and finally a relevance evaluation is derived by an intelligence module on basis of the expert knowledge in accordance with user-specific criteria.

25. The system according to claim 1, wherein extracted knowledge from the knowledge base is made available in a buffer and is reduced to relevant parts by subroutine.