Utilizing machine learning models to identify insights in a document
A device receives document information associated with a document, and receives a request to identify insights in the document information. The device performs, based on the request, natural language processing on the document information to identify words, phrases, and sentences in the document information, and utilizes a first machine learning model with the words, the phrases, and the sentences to identify information indicating abstract insights, concrete insights, and non-insights in the document. The device utilizes a second machine learning model to match the abstract insights with particular concrete insights that are different than the concrete insights, and utilizes a third machine learning model to determine particular insights based on the non-insights. The device generates an insight document that includes the concrete insights, the abstract insights matched with the particular concrete insights, and the particular insights determined based on the non-insights.
1. A method comprising:
training, by a device, a machine learning model by:
providing historical information to the machine learning model,
receiving, from the machine learning model, prediction information based on providing the historical information to the machine learning model,
updating the machine learning model based on the prediction information, and
selectively, based on whether incorrect predictions are being generated by the updated machine learning model:
providing the historical information to the updated machine learning model,
receiving other prediction information based on providing the historical information to the updated machine learning model, and
further updating the updated machine learning model based on the other prediction information;
utilizing, by the device, the trained machine learning model on a document to:
identify information indicating an abstract insight, a first concrete insight, and a non-insight in the document,
wherein the abstract insight is one or more of:
an insight that is not based on reality, a specific object, or an actual instance,
an insight that expresses a quality or characteristic apart from a specific object or an actual instance, or
a theoretical insight;
wherein the first concrete insight is one or more of:
an insight that is based on reality, a specific object, or an actual instance,
an insight pertaining to reality or an actual instance, or
an insight applied to an actual substance or thing, and
wherein the non-insight is one or more of:
an opinion,
a fact, or
a guess, and
one or more of:
match the abstract insight with a second concrete insight that is different than the first concrete insight, or
match the non-insight with a particular insight, provided in a repository, that is different from insights identified in the document;
generating, by the device, insight information that includes information indicating one or more of the first concrete insight, the abstract insight, or the particular insight; and
providing, by the device and for display, the insight information.
2. The method of claim 1 , wherein the document is a first document, the abstract insight is a first abstract insight, the non-insight is a first non-insight, the particular insight is a first particular insight, and the insight information is first insight information; and
wherein the method further comprises:
utilizing, based on one or more of past behavior information, user settings, a type of a second document, or a time of day, the machine learning model or the trained machine learning model on the second document to one or more of:
identify information indicating a second abstract insight, a third concrete insight, and a second non-insight in the second document,
match the second abstract insight with a fourth concrete insight that is different than the third concrete insight, or
match a second particular insight with the second non-insight;
generating second insight information that includes information indicating one or more of the third concrete insight, the second abstract insight, or the second particular insight; and
providing, for display, the second insight information.
3. The method of claim 1 , wherein providing, for display, the insight information comprises:
providing the insight information in the document.
4. The method of claim 1 , wherein the document is a first document; and
wherein the method further comprises:
determining a second document based on the insight information; and
providing, for display, the second document.
5. The method of claim 4 , further comprising:
comparing the insight information with one or more documents in the repository; and
wherein determining the second document comprises:
determining the second document based on comparing the insight information with the one or more documents in the repository.
6. The method of claim 1 , wherein providing the historical information to the updated machine learning model comprises:
providing, based on determining that incorrect predictions are being generated by the updated machine learning model, the historical information to the updated machine learning model.
7. The method of claim 1 , further comprising:
analyzing published content, and
creating an insight data structure based on analyzing the published content.
8. A device, comprising:
one or more memories; and
one or more processors communicatively coupled to the one or more memories, configured to:
provide historical information to a machine learning model;
receive, from the machine learning model, prediction information based on providing the historical information to the machine learning model;
update the machine learning model based on the prediction information;
when incorrect predictions are being generated by the machine learning model based on the machine learning model being updated:
provide the historical information to the machine learning model,
receive other prediction information based on providing the historical information to the machine learning model, and
update the machine learning model based on the other prediction information;
utilize, based on determining that correct predictions are being generated by the machine learning model, the machine learning model on a document to:
identify information indicating an abstract insight, a first concrete insight, and a non-insight in the document,
wherein the abstract insight is one or more of:
an insight that is not based on reality, a specific object, or an actual instance,
an insight that expresses a quality or characteristic apart from a specific object or an actual instance, or
a theoretical insight;
wherein the first concrete insight is one or more of:
an insight that is based on reality, a specific object, or an actual instance,
an insight pertaining to reality or an actual instance, or
an insight applied to an actual substance or thing, and
wherein the non-insight is one or more of:
an opinion,
a fact, or
a guess, and
one or more of:
match the abstract insight with a second concrete insight that is different than the first concrete insight, or
match the non-insight with a particular insight, provided in a repository, that is different from insights identified in the document;
generate insight information that includes information indicating one or more of the first concrete insight, the abstract insight, or the particular insight; and
provide, for display, the insight information.
9. The device of claim 8 , wherein the document is a first document, the abstract insight is a first abstract insight, the non-insight is a first non-insight, the particular insight is a first particular insight, and the insight information is first insight information; and
wherein the one or more processors are further configured to:
utilize, based on one or more of past behavior information, user settings, a type of a second document, or a time of day, the machine learning model on the second document to one or more of:
identify information indicating a second abstract insight, a third concrete insight, and a second non-insight in the second document,
match the second abstract insight with a fourth concrete insight that is different than the third concrete insight, or
match a second particular insight with the second non-insight;
generate second insight information that includes information indicating one or more of the third concrete insight, the second abstract insight, or the second particular insight; and
provide, for display, the second insight information.
10. The device of claim 8 , wherein the one or more processors, when providing, for display, the insight information, are configured to:
provide the insight information in the document.
11. The device of claim 8 , wherein the document is a first document; and
wherein the one or more processors are further configured to:
determine a second document based on the insight information; and
provide, for display, the second document.
12. The device of claim 11 , wherein the one or more processors are further configured to:
compare the insight information with one or more documents in the repository; and
wherein the one or more processors, when determining the second document, are configured to:
determine the second document based on comparing the insight information with the one or more documents in the repository.
13. The device of claim 8 , wherein the one or more processors are further configured to:
analyze published content, and
create an insight data structure based on analyzing the published content.
14. The device of claim 8 , wherein the historical information includes information associated with one or more of:
concepts included in one or more other documents,
abstract insights included in the one or more other documents,
concrete insights included in the one or more other documents, or
non-insights included in the one or more other documents.
15. A non-transitory computer-readable medium storing instructions, the instructions comprising:
one or more instructions that, when executed by one or more processors, cause the one or more processors to:
provide historical information to a machine learning model;
receive, from the machine learning model, prediction information based on providing the historical information to the machine learning model;
update the machine learning model based on the prediction information;
when incorrect predictions are being generated by the machine learning model based on the machine learning model being updated:
provide the historical information to the machine learning model,
receive other prediction information based on providing the historical information to the machine learning model, and
update the machine learning model based on the other prediction information;
utilize, based on determining that correct predictions are being generated by the machine learning model, the machine learning model on a document to:
identify information indicating an abstract insight, a first concrete insight, and a non-insight in the document,
wherein the abstract insight is one or more of:
an insight that is not based on reality, a specific object, or an actual instance,
an insight that expresses a quality or characteristic apart from a specific object or an actual instance, or
a theoretical insight;
wherein the first concrete insight is one or more of:
an insight that is based on reality, a specific object, or an actual instance,
an insight pertaining to reality or an actual instance, or
an insight applied to an actual substance or thing, and
wherein the non-insight is one or more of:
an opinion,
a fact, or
a guess, and
one or more of:
match the abstract insight with a second concrete insight that is different than the first concrete insight, or
match the non-insight with a particular insight, provided in a repository, that is different from insights identified in the document;
generate insight information that includes information indicating one or more of the first concrete insight, the abstract insight, or the particular insight; and
provide, for display, the insight information.
16. The non-transitory computer-readable medium of claim 15 , wherein the document is a first document, the abstract insight is a first abstract insight, the non-insight is a first non-insight, the particular insight is a first particular insight, and the insight information is first insight information; and
wherein the one or more instructions, when executed by the one or more processors, further cause the one or more processors to:
utilize, based on one or more of past behavior information, user settings, a type of a second document, or a time of day, the machine learning model on the second document to one or more of:
identify information indicating a second abstract insight, a third concrete insight, and a second non-insight in the second document,
match the second abstract insight with a fourth concrete insight that is different than the third concrete insight, or
match a second particular insight with the second non-insight;
generate second insight information that includes information indicating one or more of the third concrete insight, the second abstract insight, or the second particular insight; and
provide, for display, the second insight information.
17. The non-transitory computer-readable medium of claim 15 , wherein the one or more instructions, that cause the one or more processors to provide, for display, the insight information, cause the one or more processors to:
provide the insight information in the document.
18. The non-transitory computer-readable medium of claim 15 , wherein the document is a first document; and
wherein the one or more instructions, when executed by the one or more processors, further cause the one or more processors to:
determine a second document based on the insight information; and
provide, for display, the second document.
19. The non-transitory computer-readable medium of claim 18 , wherein the one or more instructions, when executed by the one or more processors, further cause the one or more processors to:
compare the insight information with one or more documents in the repository; and
wherein the one or more instructions, that cause the one or more processors to determine the second document, cause the one or more processors to:
determine the second document based on comparing the insight information with the one or more documents in the repository.
20. The non-transitory computer-readable medium of claim 15 , wherein the historical information includes information associated with one or more of:
concepts included in one or more other documents,
abstract insights included in the one or more other documents,
concrete insights included in the one or more other documents, or
non-insights included in the one or more other documents.