Training machine learning models to automatically detect and correct contextual and logical errors
Provided are techniques for training Machine Learning (ML) models to automatically detect and correct contextual and logical errors. A plurality of machine learning models are trained. In response to receiving content, the content is parsed into different elements. One or more knowledge graphs are built based on the different elements. One or more machine learning models are selected from the plurality of machine learning models based on the one or more knowledge graphs and custom criteria in a user profile. The selected one or more machine learning models are used to identify at least one of a contextual error and a logical error and a correction for the at least one of the contextual error and the logical error. The correction for the at least one of the contextual error and the logical error is applied to generate corrected content. The corrected content is rendered.
1 . A computer-implemented method, comprising operations for:
training a plurality of machine learning models; and
in response to receiving content,
building one or more knowledge graphs based on the content;
selecting a first subset of machine learning models from the plurality of machine learning models based on the one or more knowledge graphs and custom criteria in a user profile;
using the selected one or more machine learning models to identify at least one of a contextual error and a logical error and a correction for the at least one of the contextual error and the logical error;
applying the correction for the at least one of the contextual error and the logical error to generate corrected content;
rendering the corrected content;
adjusting the custom criteria based on user feedback on the corrected content:
in response to receiving new content,
adjusting the one or more knowledge graphs based on the new content; and
selecting a second subset of machine learning models from the plurality of machine learning models based on the one or more adjusted knowledge graphs and the adjusted custom criteria.
2 . The computer-implemented method of claim 1 , wherein the content is parsed into different elements based on time, location, event, subject, verb, and object.
3 . The computer-implemented method of claim 1 , further comprising operations for:
updating the one or more knowledge graphs based on receiving additional content.
4 . The computer-implemented method of claim 1 , further comprising operations for:
receiving user input to configure the custom criteria, wherein the custom criteria comprises identification of monitored applications, identification of types of errors to detect, and how to process the types of errors.
5 . The computer-implemented method of claim 1 , further comprising operations for:
generating a data structure, wherein the data structure comprises columns for: a user identifier, an application identifier, a content identifier, a content type, a content category, a topic, an input buffer, a paragraph identifier, a sentence identifier, an error type, a related sentence identifier, suggested content, and proof reading comments.
6 . The computer-implemented method of claim 1 , wherein each machine learning model of the plurality of machine learning models is retrained using user feedback on the correction.
7 . The computer-implemented method of claim 1 , wherein training data for the plurality of machine learning models comprises sample pairs of a tagged error type and a corresponding correction.
8 . A computer program product, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a processor to cause the processor to perform operations for:
training a plurality of machine learning models; and
in response to receiving content,
building one or more knowledge graphs based on the content;
selecting a first subset of machine learning models from the plurality of machine learning models based on the one or more knowledge graphs and custom criteria in a user profile;
using the selected one or more machine learning models to identify at least one of a contextual error and a logical error and a correction for the at least one of the contextual error and the logical error;
applying the correction for the at least one of the contextual error and the logical error to generate corrected content;
rendering the corrected content;
adjusting the custom criteria based on user feedback on the corrected content;
in response to receiving new content,
adjusting the one or more knowledge graphs based on the new content; and
selecting a second subset of machine learning models from the plurality of machine learning models based on the one or more adjusted knowledge graphs and the adjusted custom criteria.
9 . The computer program product of claim 8 , wherein the content is parsed into different elements based on time, location, event, subject, verb, and object.
10 . The computer program product of claim 8 , wherein the program instructions are executable by the processor to cause the processor to perform operations further comprising:
updating the one or more knowledge graphs based on receiving additional content.
11 . The computer program product of claim 8 , wherein the program instructions are executable by the processor to cause the processor to perform operations further comprising:
receiving user input to configure the custom criteria, wherein the custom criteria comprises identification of monitored applications, identification of types of errors to detect, and how to process the types of errors.
12 . The computer program product of claim 8 , wherein the program instructions are executable by the processor to cause the processor to perform operations further comprising:
generating a data structure, wherein the data structure comprises columns for: a user identifier, an application identifier, a content identifier, a content type, a content category, a topic, an input buffer, a paragraph identifier, a sentence identifier, an error type, a related sentence identifier, suggested content, and proof reading comments.
13 . The computer program product of claim 8 , wherein each machine learning model of the plurality of machine learning models is retrained using user feedback on the correction.
14 . The computer program product of claim 8 , wherein training data for the plurality of machine learning models comprises sample pairs of a tagged error type and a corresponding correction.
15 . A computer system, comprising:
one or more processors, one or more computer-readable memories and one or more computer-readable, tangible storage devices; and
program instructions, stored on at least one of the one or more computer-readable, tangible storage devices for execution by at least one of the one or more processors via at least one of the one or more computer-readable memories, to perform operations comprising:
training a plurality of machine learning models; and
in response to receiving content,
building one or more knowledge graphs based on the content;
selecting a first subset of machine learning models from the plurality of machine learning models based on the one or more knowledge graphs and custom criteria in a user profile;
using the selected one or more machine learning models to identify at least one of a contextual error and a logical error and a correction for the at least one of the contextual error and the logical error;
applying the correction for the at least one of the contextual error and the logical error to generate corrected content;
rendering the corrected content;
adjusting the custom criteria based on user feedback on the corrected content;
in response to receiving new content,
adjusting the one or more knowledge graphs based on the new content; and
selecting a second subset of machine learning models from the plurality of machine learning models based on the one or more adjusted knowledge graphs and the adjusted custom criteria.
16 . The computer system of claim 15 , wherein the content is parsed into different elements based on time, location, event, subject, verb, and object.
17 . The computer system of claim 15 , wherein the operations further comprise:
updating the one or more knowledge graphs based on receiving additional content.
18 . The computer system of claim 15 , wherein the operations further comprise:
receiving user input to configure the custom criteria, wherein the custom criteria comprises identification of monitored applications, identification of types of errors to detect, and how to process the types of errors.
19 . The computer system of claim 15 , wherein the operations further comprise:
generating a data structure, wherein the data structure comprises columns for: a user identifier, an application identifier, a content identifier, a content type, a content category, a topic, an input buffer, a paragraph identifier, a sentence identifier, an error type, a related sentence identifier, suggested content, and proof reading comments.
20 . The computer system of claim 15 , wherein each machine learning model of the plurality of machine learning models is retrained using user feedback on the correction.