Named entity (NE) interface for multiple client application programs

• US 20030130835A1
• Filed: 01/07/2002
• Published: 07/10/2003
• Est. Priority Date: 01/07/2002
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Viegao, Evelyne, Azzam, Saliha, Powell, Kevin R., Calcagno, Michael V., Weise, David N., Barklund, Par Jonas, Knoll, Sonja S.
• Granted Patent: US 7,315,810 B2
• Time in Patent Office: Days
• US Class Current: 704/1
• CPC Class Codes: G06F 40/295 Named entity recognition

Abstract

The present invention is a named entity (NE) interface to a linguistic analysis layer. The interface exposes each input sentence to the NE recognizers of all applications and returns all recognized NEs. Thus, the present invention can accommodate NEs which dynamically change in the applications, because each input string will be handed to the applications for NE recognition. The present invention also includes a data structure which is a normalized form of recognized NEs.

Claims

1. A method of linguistically analyzing a textual input, comprising:

receiving a textual representation of the textual input;

providing the textual representation to an interface that exposes the textual representation to a plurality of application programs;

receiving from the interface an indication of named entities (NEs) recognized by the application programs; and

providing NE structures representative of the recognized NEs to a linguistic analysis component.

2. The method of claim 1 wherein providing NE structures comprises:

normalizing the NE structures to a normalized form.

3. The method of claim 2 wherein the textual input comprises an input string and further comprising:

preprocessing the input string to obtain a set of tokens corresponding to the input string.

4. The method of claim 3 wherein normalizing comprises:

identifying a span of tokens covered by each of the NE structures.

5. The method of claim 4 wherein providing NE structures comprises:

identifying NE structures that cover a same span of tokens in the input string.

6. The method of claim 5 wherein providing NE structures comprises:

merging the identified NE structures into a single syntactic structure.

7. The method of claim 1 and further comprising:

providing the textual representation to a linguistic recognition component to identify NEs in the textual input in addition to those received from the interface.

8. The method of claim 7 and further comprising:

receiving NE lexicons or NE grammars from the application programs, accessible by the linguistic recognition component.

9. The method of claim 8 wherein the linguistic recognition component recognizes NEs in the textual input by accessing the NE lexicons or NE grammars.

10. The method of claim 9 wherein providing NE structures comprises:

preprocessing the textual input to obtain a set of tokens corresponding to the textual input;

normalizing the NE structures to a normalized form by identifying a span of tokens covered by each of the NE structures;

identifying NE structures that cover a same span of tokens in the textual input; and

merging the identified NE structures into a single syntactic structure.

11. A linguistic analysis system receiving a plurality of textual inputs and providing an output, comprising:

an interface exposing each of the textual inputs to a plurality of external named entity (NE) recognizers each associated with an application program; and

an NE structure generator generating NE structures representative of recognized NEs, recognized by the external NE recognizers.

12. The linguistic analysis system of claim 11 and further comprising:

a linguistic layer recognition component configured to receive the textual input and recognize NEs in addition to the NEs recognized by the external NE recognizers.

13. The linguistic analysis system of claim 12 wherein the linguistic layer recognition component is configured to receive at least one of NE lexicons and NE grammars from the application programs and recognize NEs by accessing the NE grammars and NE lexicons.

14. The linguistic analysis system of claim 11 wherein each external NE recognizer is associated with an application program and wherein the interface comprises:

an application programming interface which, when invoked, exposes a textual input to the external NE recognizers and receives the recognized NEs from the external recognizers along with indications of the application programs associated with each external NE recognizer that recognized a NE.

15. The linguistic analysis system of claim 12 wherein the NE structure generator comprises:

a normalizing component configured to normalize the recognized NEs to a normalized form.

16. The linguistic analysis system of claim 15 wherein the textual input comprises an input string and further comprising:

a preprocessing component configured to preprocess the input string to obtain a set of tokens corresponding to the input string.

17. The linguistic analysis system of claim 16 wherein the normalizing component comprises:

a merging component identifying a span of tokens covered by each of the NE structures.

18. The linguistic analysis system of claim 17 wherein the merging component is further configured to identifying NE structures that cover a same span of tokens in the input string.

19. The linguistic analysis component of claim 18 wherein the merging component is configured to merge the identified NE structures into a single syntactic structure.

20. A linguistic analysis system providing a textual input to a plurality of named entity (NE) recognizers, each NE recognizer being associated with an application program and identifying NEs in the textual input, the linguistic analysis system providing a data structure indicative of each of the identified NEs, the data structure comprising:

an application identifier portion identifying the application associated with the NE recognizer that recognized the NE;

a semantic type portion identifying a semantic type of the recognized NE; and

a span portion indicative of a span of the textual input covered by the recognized NE.

21. The linguistic analysis system of claim 20 wherein the data structure further comprises:

a NE identifier portion identifying the recognized NE.

22. The linguistic analysis system of claim 21 wherein the data structure further comprises:

a syntactic portion identifying a syntactic type of the recognized NE.

23. The linguistic analysis system of claim 21 wherein the data structure further comprises:

an attributes portion identifying a data structure comprising the recognized NE.

24. The linguistic analysis system of claim 23 wherein the data structure further comprises:

an NE string portion identifying a string of the recognized NE.

25. The linguistic analysis system of claim 23 wherein the span portion comprises:

a first token identifier identifying the first token in the input string covered by the recognized NE.

26. The linguistic analysis system of claim 25 wherein the span portion comprises:

an ending token identifier identifying an ending token in the input string covered by the recognized NE.