Knowledge base for analysis of text

• US 10,679,008 B2
• Filed: 04/14/2017
• Issued: 06/09/2020
• Est. Priority Date: 12/16/2016
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Inventors: Dubey, Rajeev, Jain, Navendu
• Primary Examiner(s): Kim, Jonathan C
• Application Number:
  US15/487,960
• Publication Number:
  US 20180173698A1
• Time in Patent Office: 1,152 Days
• Field of Search: None
• CPC Class Codes: G06F 16/3347 using vector based model
• CPC Class Codes: G06F 40/30 Semantic analysis

Abstract

A knowledge base can include a dictionary associated with classes of a model, e.g., an ontology. A text segment that is not found in the dictionary can be received. Feature(s) can be determined for the text segment and, based partly on providing the feature(s) to a classifier, a set of values can be determined. The distribution can include values respectively corresponding to the classes. One of the values can be greater than a predetermined threshold. That value can correspond to a class. An indication identifying the class can be presented via a user interface having functionality to provide input that the text segment is associated with the class, is not associated with the class, or is associated with another class. Based at least partly on adding a new class to the ontology, a precedence table indicating priorities between motifs defining relationships between classes of the ontology can be updated.

Claims

1. A system comprising:

at least one processor;

memory communicatively coupled to the at least one processor; and

instructions stored in the memory and executable by the at least one processor to;

receive, from a computing device operated by a user, a document associated with a first text segment and a second text segment, wherein the document includes a user experience written in natural language;

identify a motif for the first text segment and the second text segment based on an inferred meaning of the first text segment and the second text segment;

access a predetermined knowledge base including a dictionary associated with an ontology comprising a plurality of classes;

determine a first set of values associated with the first text segment, the first set of values including a plurality of first values respectively corresponding to individual classes of the plurality of classes;

determine that a first value of the plurality of first values is greater than a first predetermined threshold, the first value corresponding to a first class of the plurality of classes;

infer a first actionable item associated with the first text segment based at least in part on the first class and the motif, wherein the first actionable item is an instruction for an action to be performed by another user to respond to the first text segment according to the motif, with the action corresponding to the first class and the motif;

infer a second actionable item associated with the second text segment;

determine a priority for the first actionable item and the second actionable item based on the motifs;

determine a ranked list of actionable items based on the priority for the first actionable item and the priority for the second actionable item; and

transmit the ranked list of actionable items to a computing device of the another user.

2. The system as claim 1 recites, wherein the one or more modules stored in the memory are further executable by the at least one processor to:

determine a feature vector associated with the first text segment, wherein the feature vector comprises at least one of a parts-of-speech tag for the first text segment, stemmed output associated with the first text segment in view of other stemmed output, or a co-occurrence relationship between the first text segment and second text segments used in a similar context as the first text segment; and

determine the first set of values based on the feature vector.

3. The system as claim 1 recites, wherein the plurality of classes of the ontology comprise at least one of a suggestion-indicator class for text segment(s) that represent the presence of a suggestion, a modal-indicator class for text segment(s) that represent the presence of a suggestion, an entity class for text segment(s) that represents an entity, an action class for text segment(s) that represents a behavior that is taken on an entity, a certainty class for text segment(s) that represents the confidence of the user regarding an entity or condition represented in the text segment(s), an emotion class for text segment(s) that represents feelings or beliefs of the user writing the text segment(s), a negation class for text segment(s) that change the meaning of other text segment(s) to a substantially opposite meaning, or a qualifier class for text segment(s) that represents descriptive terms used to describe an action or a condition.

4. The system as claim 1 recites, wherein the instructions are further executable by the at least one processor to:

determine a second set of values associated with the second text segment, the second set of values including a plurality of second values respectively corresponding to individual classes of the plurality of classes;

determine that each second value of the plurality of second values is less than or equal to the first predetermined threshold;

determine a highest second value of the plurality of second values, the highest second value corresponding to a second class of the plurality of classes;

wherein the inference of the second actionable item associated with the second text segment is based at least in part on the second class, wherein the second actionable item is an instruction for an action to be performed by the another user, with the action corresponding to the second class.

5. The system as claim 1 recites, wherein the instructions are further executable by the at least one processor to:

determine a second set of values associated with the second text segment, the second set of values including a plurality of second values respectively corresponding to individual classes of the plurality of classes;

determine that a second value of the plurality of second values is greater than a second predetermined threshold, the second value corresponding to a third class of the plurality of classes; and

automatically map the second text segment to a second class of the plurality of classes.

6. The system as claim 5 recites, wherein the second predetermined threshold is associated with a greater value than the first predetermined threshold.

7. The system as claim 1 recites, wherein the instructions are further executable by the at least one processor to:

determine a second set of values associated with the second text segment, the second set of values including a plurality of second values respectively corresponding to individual classes of the plurality of classes;

determine that a second value of the plurality of second values is less than a second predetermined threshold, the second value corresponding to a second class of the plurality of classes; and

automatically refrain from mapping the second text segment to a second class of the plurality of classes.

8. The system as claim 1 recites, wherein the instructions are further executable by the at least one processor to:

receive, from a computing device operated by a second user, a new text segment;

determine, based at least in part on the dictionary, that the new text segment is associated with a second class of the plurality of classes; and

infer an actionable item associated with the new text segment based at least in part on the second class and at least one grammar pattern.

9. A computer-implemented method comprising:

receiving, from a computing device operated by a user, a document associated with a first text segment and a second text segment, wherein the document includes a user experience written in natural language;

accessing a predetermined knowledge base including;

a dictionary associated with an ontology comprising a plurality of classes; and

a precedence table, the precedence table establishing a priority between motifs;

identify a set of classes for the first text segment;

determining, based at least in part on the dictionary and the set of classes, a first motif and a second motif for the first text segment, the first motif defining a first relationship between individual classes of the plurality of classes and the second motif defining a second relationship between individual classes of the plurality of classes;

determining, based at least in part on the precedence table, that the first motif precedes the second motif, wherein the precedence table ranks a plurality of motifs by priority;

inferring a first actionable item associated with the first text segment based at least in part on the first motif, wherein the first actionable item is an instruction for an action to be performed by another user to respond to the first text segment according to the first motif, with the action corresponding to the first motif;

inferring a second actionable item associated with the second text segment;

determining a priority for the first actionable item and the second actionable item based on the motifs;

determining a ranked list of actionable items based on the priority for the first actionable item and the priority for the second actionable item; and

transmitting the ranked list of actionable items to a computing device of the another user.

10. The computer-implemented method as claim 9 recites, further comprising:

determining, based at least in part on the dictionary, that the first text segment is associated with a third motif defining a third relationship between individual classes of the plurality of classes; and

determining, based at least in part on the precedence table, that the first motif precedes the third motif and that the third motif precedes the second motif.

11. The computer-implemented method as claim 9 recites, further comprising:

determining, based at least in part on the dictionary, that the first text segment is associated with a third motif defining a third relationship between individual classes of the plurality of classes; and

determining, based at least in part on the precedence table, that the second motif precedes the third motif.

12. The computer-implemented method as claim 9 recites, further comprising, prior to receiving the document associated with the first text segment and the second text segment, generating the precedence table based at least in part on:

associating individual motifs with values representative of importance; and

determining prioritizations between the individual motifs based at least in part on the values.

13. The computer-implemented method as claim 9 recites, further comprising:

updating the dictionary based at least in part on;

receiving a new text segment;

determining a first set of values associated with the new text segment, the first set of values including a plurality of first values respectively corresponding to individual classes of the plurality of classes;

determining that a first value of the plurality of first values is greater than a first predetermined threshold, the first value corresponding to a particular class of the plurality of classes;

presenting, via a user interface, a first indication identifying the particular class, the user interface configured to receive at least one of;

a first command to cause the new text segment to be mapped to a first entry in the dictionary corresponding to the particular class;

or

a second command to cause the new text segment to be mapped to a second entry in the dictionary corresponding to a different class of the plurality of classes instead of the first entry; and

receiving input corresponding to the second command; and

updating the precedence table based at least in part on the input.

14. One or more non-transitory computer readable media storing computer-executable instructions that, when executed by the at least one processor, cause the at least one processor to perform acts comprising:

receiving, from a computing device operated by a user, a document associated with a first text segment and a second text segment, wherein the document includes a user experience written in natural language;

identifying a motif for the first text segment and the second text segment based on an inferred meaning of the first text segment and the second text segment;

accessing a predetermined knowledge base including a dictionary associated with an ontology comprising a plurality of classes, a class of the plurality of classes being associated with a positive sub-class, a negative sub-class, and a neutral-subclass;

determining, based at least in part on the ontology, that the first text segment and the second text segment are associated with the positive sub-class, the negative sub-class, or the neutral-subclass;

inferring a first actionable item associated with the first text segment based at least in part on the positive sub-class, the negative sub-class, or the neutral-subclass, the motif, and at least one grammar pattern, wherein the first actionable item is an instruction for an action to be performed by another user to respond to the first text segment according to the motif, with the action corresponding to the positive sub-class, the negative sub-class, or the neutral-subclass and the motif;

inferring a second actionable item associated with the second text segment;

determining a priority for the first actionable item and the second actionable item based on the motifs;

determining a ranked list of actionable items based on the priority for the first actionable item and the priority for the second actionable item; and

transmitting the ranked list of actionable items to a computing device of the another user.

15. The one or more non-transitory computer readable media as claim 14 recites, the acts further comprising updating the dictionary based at least in part on:

receiving a new text segment;

determining a first set of values associated with the new text segment, the first set of values including a plurality of first values respectively corresponding to individual classes of the plurality of classes;

determining that a first value of the plurality of first values is greater than a first predetermined threshold, the first value corresponding to a particular class of the plurality of classes;

presenting, via a user interface, a first indication identifying the particular class, the user interface configured to receive at least one of;

a first command to cause the first text segment to be mapped to a first entry in the dictionary corresponding to the particular class;

or

a second command to cause the first text segment to be mapped to a second entry in the dictionary corresponding to a different class of the plurality of classes instead of the first entry; and

receiving input corresponding to the first command or the second command.