EMAIL-LIKE USER INTERFACE FOR TRAINING NATURAL LANGUAGE SYSTEMS

• US 20150302850A1
• Filed: 04/14/2015
• Published: 10/22/2015
• Est. Priority Date: 04/16/2014
• Status: Active Grant

• Current Assignee: Meta Platforms, Inc. (f/k/a Facebook, Inc.)
• Original Assignee: Meta Platforms, Inc. (f/k/a Facebook, Inc.)
• Inventors: Lebrun, Alexandre
• Granted Patent: US 10,978,052 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 40/35 Discourse or dialogue representation
• CPC Class Codes: G10L 15/063 Training
• CPC Class Codes: G10L 15/18 using natural language modelling

Abstract

An email-like user interface displays a list of user logs determined based on user-specified list criteria to user logs received in a natural language (NL) training environment. The list comprise a subset of the received user logs in order to minimize the number of actions required to configure and train the NL configuration system in a semi-supervised manner, thereby improving the quality and accuracy of NL configuration system. To determine a list of user logs relevant for training the user logs can be filtered, sorted, grouped and searched within the email-like user interface. A training interface to a network of instances that comprises a plurality of NL configuration systems leverages a crowd-sourcing community of developers in order to efficiently create a customizable NL configuration system.

Claims

1. A computer-implemented method comprising:

receiving a plurality of user logs from a plurality of runtime systems or applications, each user log comprising a natural language expression and an intent, the intent being predicted by the natural language configuration system based on the natural language expression;

determining at least one of a filtered, grouped, or sorted list of user logs, the list comprising one or more of the received user logs;

displaying a plurality of action panels in an inbox view, each action panel being associated with one or more different user log of the list of user logs and comprising the natural language expression and the intent of the associated user log;

displaying for each action panel an option to validate or dismiss the associated user log; and

in response to receiving a selection of the option to validate the user log, configuring and training the natural language configuration system based on the natural language expression and the intent of the validated user log.

2. The method of claim 1, wherein each action panel further comprises a confidence score calculated by the natural language configuration system when predicting the intent based on the natural language expression.

3. The method of claim 1, wherein the list of user logs is determined by filtering the received plurality of user logs based on a filter criterion and list criteria comprise the filter criterion.

4. The method of claim 1, wherein the list of user logs is determined by grouping the received plurality of user logs based on attributes of the user logs and the list criteria comprise the attributes of the user logs.

5. The method of claim 1, wherein the list of user logs is sorted based on attributes of the user logs and the list criteria comprise the attributes of the user logs.

6. The method of claim 1, wherein associating each action panel with a different user log is in the order of the different user logs within the list of user logs.

7. The method of claim 1, wherein one or more user logs from the plurality of received user logs were created by instances of natural language configuration systems included in network of instances.

8. The method of claim 7, wherein one or more user logs from the plurality of received user logs comprises an update to a forked instance of the natural language configuration system.

9. A computer program product comprising a non-transitory computer readable storage medium for version control for application development, the non-transitory computer readable storage medium storing instructions for:

receiving a plurality of user logs from a plurality of runtime systems or applications, each user log comprising a natural language expression and an intent, the intent being predicted by the natural language configuration system based on the natural language expression;

determining at least one of a filtered, grouped, or sorted list of user logs, the list comprising one or more of the received user logs;

displaying a plurality of action panels in an inbox view, each action panel being associated with one or more different user log of the list of user logs and comprising the natural language expression and the intent of the associated user log;

displaying for each action panel an option to validate or dismiss the associated user log; and

in response to receiving a selection of the option to validate the user log, configuring and training the natural language configuration system based on the natural language expression and the intent of the validated user log.

10. The computer program product of claim 9, wherein each action panel further comprises a confidence score calculated by the natural language configuration system when predicting the intent based on the natural language expression.

11. The computer program product of claim 9, wherein the list of user logs is determined by filtering the received plurality of user logs based on a filter criterion and list criteria comprise the filter criterion.

12. The computer program product of claim 9, wherein the list of user logs is determined by grouping the received plurality of user logs based on attributes of the user logs and the list criteria comprise the attributes of the user logs.

13. The computer program product of claim 9, wherein the list of user logs is sorted based on attributes of the user logs and the list criteria comprise the attributes of the user logs.

14. The computer program product of claim 9, wherein associating each action panel with a different user log is in the order of the different user logs within the list of user logs.

15. The computer program product of claim 9, wherein one or more user logs from the plurality of received user logs were created by instances of natural language configuration systems included in network of instances.

16. The computer program product of claim 15, wherein one or more user logs from the plurality of received user logs comprises an update to a forked instance of the natural language configuration system.