Systems and methods for managing a complex lexicon comprising multiword expressions and multiword inflection templates

• US 8,762,131 B1
• Filed: 06/17/2009
• Issued: 06/24/2014
• Est. Priority Date: 06/17/2009
• Status: Active Grant

• Current Assignee: Softwin SRL Romania
• Original Assignee: Softwin SRL Romania
• Inventors: Diaconescu, Stefan Stelian, Mateescu, Marius, Minca, Andrei, Masei, Gabriel, Paun, Bianca-Daniela
• Primary Examiner(s): Armstrong, Angela A
• Application Number:
  US12/486,629
• Time in Patent Office: 1,833 Days
• Field of Search: 704 1- 10
• US Class Current: 704/9
• CPC Class Codes: G06F 40/242 Dictionaries

Abstract

In some embodiments, linguistic knowledge bases including multiword lexicon entries belonging to various natural languages are generated and maintained. Multiword lexicon entries include representations of multiword expressions (e.g. composite words such as “expectation value”, verbal phrases such as “to stumble upon”, and idiomatic phrases such as “to hit the nail on the head”), and multiword inflection templates (e.g. “more adjective” as a symbolic representation for the comparative form of adjectives). A representation of a multiword lexicon entry, including a dependency tree and/or an attribute-value tree, is displayed within a graphical user interface, allowing a user to verify, update, and correct lexical data. Computer-readable encodings (e.g. XML) of dependency tree and/or attribute-value tree data are saved as part of each lexicon entry.

Claims

1. A system comprising at least one processor configured to form:

a computer-implemented lexicon manager configured to

receive a multiword lexicon entry for a multiword expression of a natural language, wherein the multiword lexicon entry comprises Grammar Abstract Language (GAL) code describing a dependency tree comprising a plurality of nodes, the GAL code describing an attribute-value tree for each dependency tree node, and

in response to displaying a graphical representation of the dependency tree on a computer-implemented display device, receive a user input indicating a modification of the dependency tree, and determine an edited lexicon entry according to the multiword lexicon entry and the modification, wherein the edited lexicon entry comprises GAL code, and wherein the display device is configured to receive user input indicating modifications of the dependency tree including a dependency tree node creation, a dependency tree node deletion, an inter-node link creation, an inter-node link deletion, a modification of a type of inter-node link, and a textual annotation of a dependency tree node; and

a GAL compiler connected to the lexicon manager and configured to compile the GAL code representing the edited lexicon entry into Extensible Markup Language (XML) code.

2. The system of claim 1, wherein the GAL compiler is further configured, in response to compiling the GAL code, to transmit the XML code to a lexicon knowledgebase stored on a computer-readable medium.

3. The system of claim 1, wherein the display device is configured to receive user input indicating modifications of an attribute value tree including replacing a value of an attribute node of the attribute value tree and providing a missing value to an attribute node of the attribute value tree.

4. A computer-implemented method comprising employing at least one processor to perform the steps of:

receiving a multiword lexicon entry for a multiword expression of a natural language, wherein the multiword lexicon entry comprises Grammar Abstract Language (GAL) code describing a dependency tree comprising a plurality of nodes, the GAL code describing an attribute-value tree for each dependency tree node;

in response to displaying a graphical representation of the dependency tree on a computer-implemented display device, receiving a user input indicating a modification of the dependency tree, and determining an edited lexicon entry according to the multiword lexicon entry and the modification, wherein the edited lexicon entry comprises GAL code, and wherein the display device is configured to receive user input indicating modifications of the dependency tree including a dependency tree node creation, a dependency tree node deletion, an inter-node link creation, an inter-node link deletion, a modification of a type of inter-node link, and a textual annotation of a dependency tree node; and

compiling the GAL code representing the edited lexicon entry into Extensible Markup Language (XML) code.

5. The system of claim 4, further comprising, in response to compiling the GAL code, transmitting the XML code to a lexicon knowledgebase stored on a computer-readable medium.

6. The method of claim 4, wherein the display device is configured to receive user input indicating modifications of an attribute value tree including replacing a value of an attribute node of the attribute value tree and providing a missing value to an attribute node of the attribute value tree.

7. A system comprising at least one processor configured to form:

a computer-implemented lexicon manager configured to

receive a multiword expression formulated in a natural language,

perform a syntactic analysis of the multiword expression to produce a dependency tree of the multiword expression, the dependency tree comprising a plurality of nodes,

formulate a lexicon entry of the multiword expression in Grammar Abstract Language (GAL) code, the lexicon entry comprising the dependency tree, the GAL code describing an attribute-value tree for each dependency tree node, and

in response to displaying a graphical representation of the dependency tree on a computer-implemented display device, receive a user input indicating a modification of the dependency tree, and determine an edited lexicon entry according to the multiword lexicon entry and the modification, wherein the edited lexicon entry comprises GAL code, and wherein the display device is configured to receive user input indicating modifications of the dependency tree including a dependency tree node creation, a dependency tree node deletion, an inter-node link creation, an inter-node link deletion, a modification of a type of inter-node link, and a textual annotation of a dependency tree node; and

a GAL compiler connected to the lexicon manager and configured to compile the GAL representing the edited lexicon entry code into Extensible Markup Language (XML) code.

8. The system of claim 7, wherein the GAL compiler is further configured, in response to compiling the GAL code, to transmit the XML code to a lexicon knowledgebase stored on a computer-readable medium.

9. The system of claim 7, wherein the display device is configured to receive user input indicating modifications of an attribute value tree including replacing a value of an attribute node of the attribute value tree and providing a missing value to an attribute node of the attribute value tree.

10. A computer-implemented method comprising employing at least one processor to:

perform a syntactic analysis of a multiword expression formulated in a natural language to produce a dependency tree of the multiword expression, the dependency tree including a plurality of nodes;

formulate a lexicon entry of the multiword expression in Grammar Abstract Language (GAL) code, the lexicon entry comprising the dependency tree, the GAL code describing an attribute-value tree for each dependency tree node;

in response to displaying a graphical representation of the dependency tree on a computer-implemented display device, receive a user input indicating a modification of the dependency tree, and determine an edited lexicon entry according to the multiword lexicon entry and the modification, wherein the edited lexicon entry comprises GAL code, and wherein the display device is configured to receive user input indicating modifications of the dependency tree including a dependency tree node creation, a dependency tree node deletion, an inter-node link creation, an inter-node link deletion, a modification of a type of inter-node link, and a textual annotation of a dependency tree node; and

compile the GAL code representing the edited lexicon entry into Extensible Markup Language (XML) code.

11. The method of claim 10, further comprising, in response to compiling GAL code, transmitting the XML code to a lexicon knowledgebase stored on a computer-readable medium.

12. The method of claim 10, wherein the display device is configured to receive user input indicating modifications of an attribute value tree including replacing a value of an attribute node of the attribute value tree and providing a missing value to an attribute node of the attribute value tree.

13. A system comprising at least one processor configured to form:

means for receiving a multiword lexicon entry for a multiword expression of a natural language, wherein the multiword lexicon entry comprises Grammar Abstract Language (GAL) code describing a dependency tree comprising a plurality of nodes, the GAL code describing an attribute-value tree for each dependency tree node;

means for displaying to a user a graphical representation of the dependency tree in response to receiving the multiword lexicon entry;

means for receiving a user input indicating a modification of the dependency tree in response to displaying the dependency tree, wherein the means for receiving the user input is configured to receive user input indicating modifications of the dependency tree including a dependency tree node creation, a dependency tree node deletion, an inter-node link creation, an inter-node link deletion, a modification of a type of inter-node link, and a textual annotation of a dependency tree node;

means for determining an edited lexicon entry according to the multiword lexicon entry and the modification, wherein the edited lexicon entry comprises GAL code; and

means for compiling the GAL representing the edited lexicon entry code into Extensible Markup Language (XML) code.

14. The system of 13, further comprising means for receiving a user input indicating modifications of an attribute value tree including replacing a value of an attribute node of the attribute value tree and providing a missing value to an attribute node of the attribute value tree.