Visual display of semantic information

• US 9,524,291 B2
• Filed: 10/06/2010
• Issued: 12/20/2016
• Est. Priority Date: 10/06/2010
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: VirtuOz SA (Microsoft Corporation)
• Inventors: Teodosiu, Dan, Powers, Elizabeth Ireland, LeRoy, Pierre Serge Vincent, Saunier, Sebastien Jean-Marie Christian
• Primary Examiner(s): Desir, Pierre-Louis
• Assistant Examiner(s): Kovacek, David
• Application Number:
  US12/899,016
• Publication Number:
  US 20120089394A1
• Time in Patent Office: 2,267 Days
• Field of Search: 704 1- 8, 704/9, 704/10, 704/200, 704/231, 704236-242, 704243-245, 704250-255, 704/270, 704/276, 704E15001-E1505
• US Class Current: 1/1
• CPC Class Codes: G06F 40/00 Handling natural language data speech analysis or synthesis, speech recognition G10L
• CPC Class Codes: G06F 40/10 Text processing natural language analysis G06F40/20; semantic analysis G06F40/30; processing or translation of natural language G06F40/40
• CPC Class Codes: G06F 40/237 Lexical tools
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: G10L 21/06 Transformation of speech into a non-audible representation, e.g. speech visualisation or speech processing for tactile aids G10L15/26 takes precedence

Abstract

Techniques involving visual display of information related to matching user utterances against graph patterns are described. In one or more implementations, an utterance of a user is obtained that has been indicated as corresponding to a graph pattern through linguistic analysis. The utterance is displayed in a user interface as a representation of the graph pattern.

Claims

1. A method implemented by one or more computer processors, the method comprising:

generating, through a linguistic analysis, a semantic graph of a user utterance received by a conversational agent implementing an automated interactive natural language conversational dialog with a user, the conversational dialog modeled as paths through a set of decision trees;

identifying a plurality of candidate graph patterns based upon a current position of the conversational dialog in the set of decision trees;

comparing through use of one or more distance measuring metrics, the semantic graph of the user utterance with the plurality of candidate graph patterns to determine a given graph pattern from among the plurality of candidate graph patterns corresponding to the user utterance, the given graph pattern being a matching graph pattern and comprising a plurality of graph fragments, wherein each graph fragment occurs in the semantic graph of the user utterance or the matching graph pattern being subsumed by a subgraph of the semantic graph of the user utterance; and

displaying the utterance in an electronic user interface as a representation of the matching graph pattern, the displayed utterance facilitating improvement to the conversational agent.

2. A method as described in claim 1, wherein the displaying includes displaying a control that is selectable to expand the representation to include another utterance that has been indicated as corresponding to the matching graph pattern through linguistic analysis.

3. A method as described in claim 2, wherein the displaying of the utterance and the other utterance are ordered by specificity in the user interface.

4. A method as described in claim 1, wherein the displaying includes displaying an indication that describes popularity of the matching graph pattern in a corpus of utterances that was a subject of the linguistic analysis.

5. A method as described in claim 1, wherein the displaying includes displaying in the user interface a link that is selectable to navigate to a conversation that includes the utterance.

6. A method as described in claim 1, wherein the displaying is performed to indicate which words in the utterance served as a basis to determine that the utterance corresponds to the matching graph pattern.

7. A method as described in claim 1, wherein the displaying is performed to indicate which words in the utterance are not included in a lexicon that is utilized to perform the linguistic analysis.

8. A method as described in claim 7, wherein at least one of the words is a misspelling.

9. A method as described in claim 1, wherein the displaying is performed for a plurality of said utterances that correspond to a respective one of a plurality of said graph patterns of an intent.

10. A method as described in claim 1, wherein the interactive natural language dialog is a textual chat between the user and the conversational agent.

11. A method as described in claim 1, wherein comparing, through use of the one or more distance measuring metrics, the semantic graph of the user utterance with the plurality of candidate graph patterns includes:

assigning a matching distance to each successful match between the semantic graph of the user utterance and a respective one of the plurality of candidate graph patterns; and

selecting the matching graph pattern as a respective one of the plurality of candidate graph patterns with a smallest matching distance.

12. A method as described in claim 1, wherein comparing, through use of the one or more distance measuring metrics, the semantic graph of the user utterance with the plurality of candidate graph patterns includes measuring a distance between the semantic graph of the user utterance and at least one of the plurality of candidate graph patterns.

13. A data processing system including one or more processors, the system comprising:

a conversational agent, executing on the one or more processors, the conversational agent configured to process a user utterance received by the conversational agent while implementing an automated interactive natural language conversational dialog between the conversational agent and a user, the conversational dialog modeled as paths through a set of decision trees and the conversational agent configured to;

generate, through a linguistic analysis, a semantic graph of the user utterance;

identify a plurality of candidate graph patterns based upon a current position of the conversational dialog in the set of decision trees;

compare, through use of one or more distance measuring metrics, the semantic graph of the user utterance with the plurality of candidate graph patterns to determine a given graph pattern from among the plurality of candidate graph patterns corresponding to the user utterance, the given graph pattern being a matching graph pattern and comprising a plurality of graph fragments, wherein each graph fragment occurs in the semantic graph of the user utterance or the matching graph pattern being subsumed by a subgraph of the semantic graph of the user utterance; and

a display configured to display the utterance in a user interface as a representation of the matching graph pattern, the displayed utterance facilitating improvement to the conversational agent.

14. A data processing system as described in claim 13, wherein the display is further configured to display an indication that describes popularity of the matching graph pattern in a corpus of utterances that was a subject of the linguistic analysis.

15. A data processing system as described in claim 13, wherein the display is further configured to display in the user interface a link that is selectable to navigate to a conversation that includes the utterance.

16. A data processing system as described in claim 13, wherein the display is further configured to indicate which words in the utterance are not included in a lexicon that is utilized to perform the linguistic analysis.

17. A data processing system as described in claim 13, wherein at least one of the words in the user utterance, received in connection with the dialog with a user, is a misspelling.