Personalized content distribution

1. A system comprising:

memory comprising:

a content database comprising content, associated with a course, a class, or a group within the course or the class, for delivery to a user;

a task database comprising data identifying a plurality of tasks; and

a user profile database comprising information identifying one of several attributes of the user;

a user device comprising:

a first network interface configured to exchange data via a communication network; and

a first I/O subsystem configured to convert electrical signals to user interpretable outputs via a user interface; and

one or several servers, wherein the one or several servers are configured to:

receive content corresponding to at least one source document comprising a textbook created in, or scanned into electronic format;

parse the content;

identify segments, comprising a plurality of words or phrases within a glossary or an index of the textbook, from the parsed content;

generate a networked grouping of the segments, comprising the words or phrases that are common within a table of contents, a chapter, a section, or a sub-section within the textbook;

receive historical user information about a plurality of users from the user profile database;

train a model by using the historical user information, wherein the historical user information includes interaction data associated with interactions between the plurality of users and content in the content database, the interaction data including at least one of an amount of content consumed, time spent interacting with the content, and a skill level associated with a user in the plurality of users of the content;

receive activities of the user, wherein the activities are associated with new interactions between the user and content in the content database;

parse the activities of the user;

identify components from the parsed activities;

correlate the components in the parsed activities of the user with the segments identified in the parsed content;

extract features from the activities of the user based on the correlation; and

use the trained model to estimate a mastery level, associated with the segments from the parsed content or a task in the plurality of tasks, of the user based on the features.

2. The system of claim 1 , wherein the segments comprise at least one of portions of words, words, phrases, sentences, paragraphs, images, video, or sections of the content.

3. The system of claim 1 , wherein the one or several servers are further configured to generate the networked grouping of the segments based on a table of contents of the at least one source document or HTML tags of the at least one source document.

4. The system of claim 1 , wherein the networked grouping comprises an acyclic directed network of the segments.

5. The system of claim 1 , wherein the one or several servers are further configured to train the model by:

deriving features from the historical user information;

deriving a labeled data set from the historical user information, wherein labels of the labeled data set correspond to outcomes associated with the features;

inputting a first portion of the features and a first portion of the labeled data set into the model; and

adjusting the model until correct outcomes are generated for the first portion of the features.

6. The system of claim 5 , wherein the one or several servers are further configured to train the model by:

inputting a second portion of the features and a second portion of the labeled data set into the model;

using the model to generate predicted outcomes based on the second portion of the features; and

evaluating the model by comparing the predicted outcomes with outcomes associated with the second portion of the features.

7. The system of claim 1 , wherein the model is a random forest classifier.

8. The system of claim 1 , wherein the one or several servers are further configured to train a plurality of models corresponding to different segments of the parsed content.

9. The system of claim 1 , wherein the mastery level of the user is estimated for each of a plurality of skills.

10. The system of claim 1 , wherein the features are extracted based on a position of a correlated segment within the networked grouping.

11. A method comprising:

receiving content, associated with a course, a class, or a group within the course or the class, corresponding to at least one source document comprising a textbook created in, or scanned into electronic format;

parsing the content;

identifying segments, comprising a plurality of words or phrases within a glossary or an index of the textbook, from the parsed content;

generating a networked grouping of the segments, comprising the words or phrases that are common within a table of contents, a chapter, a section, or a sub-section within the textbook;

receiving historical user information about a plurality of users;

training a model by using the historical user information, wherein the historical user information includes interaction data associated with interactions between the plurality of users and content in the content database, the interaction data including at least one of an amount of content consumed, time spent interacting with the content, and a skill level associated with a user in the plurality of users of the content;

receiving activities of a user, wherein the activities are associated with new interactions between the user and content in the content database;

parsing the activities of the user;

identifying components from the parsed activities;

correlating the components in the parsed activities of the user with the segments identified in the parsed content;

extracting features from the activities of the user based on the correlation; and

using the trained model to estimate a mastery level, associated with the segments from the parsed content or a task in the plurality of tasks, of the user based on the features.

12. The method of claim 11 , wherein the segments comprise at least one of portions of words, words, phrases, sentences, paragraphs, images, video, or sections of the content.

13. The method of claim 11 , wherein the networked grouping of the segments is generated based on a table of contents of the at least one source document or HTML, tags of the at least one source document.

14. The method of claim 11 , wherein the networked grouping comprises an acyclic directed network of the segments.

15. The method of claim 11 , wherein training the model comprises:

deriving features from the historical user information;

deriving a labeled data set from the historical user information, wherein labels of the labeled data set correspond to outcomes associated with the features;

inputting a first portion of the features into the model; and

adjusting the model until correct outcomes are generated for the first portion of features.

16. The method of claim 15 , wherein training the model further comprises:

inputting a second portion of the features and a second portion of the labeled data set into the model;

using the model to generate predicted outcomes based on the second portion of the features; and

evaluating the model by comparing the predicted outcomes with outcomes associated with the second portion of the features.

17. The method of claim 11 , wherein the model is a random forest classifier.

18. The method of claim 11 , further comprising training a plurality of models corresponding to different segments of the parsed content.

19. The method of claim 11 , wherein the mastery level of the user is estimated for each of a plurality of skills.

20. The method of claim 11 , wherein the features are extracted based on a position of a correlated segment within the networked grouping.