Electronic message information retrieval system

1. A computer system for parsing bulk message information into intelligent question retrieval models, comprising:

one or more processors; and

one or more computer-readable media having stored thereon executable instructions that when executed by the one or more processors configure the computer system to perform at least the following:

receive text-based data associated with a particular user;

parse a word from the text-based data;

encode the word into a context-dependent vector, wherein the context-dependent vector indicates the meaning of the word across a semantic space based upon the context of the word with the text-based data;

identify within a context-independent database a context-independent vector that is associated with the word, wherein the context-independent vector is based upon definitions of the word and without insight into the text-based data;

generate an objective output by combining the context-dependent vector and the context-independent vector and a regularization function; and

generate a sentence encoding representation by processing at least a portion of the text-based data through a high-level feature embedded convolutional semantic model to generate numerical representations of questions and answers within the text-based dataset, wherein the sentence encoding representation is generated at least in part based upon the objective output.

2. The computer system of claim 1 , wherein the executable instructions include instructions that are executable to configure the computer system to:

parse words from the text-based data; and

identify adjacency pairs within the words.

3. The computer system of claim 2 , wherein identifying adjacency pairs comprises:

identifying the presence of a portion of the parsed words within a first entry in an adjacency pair database; and

identifying the presence of another portion of the parsed words within a second entry in the adjacency pair database, wherein the second entry to indicated as being complimentary to the first entry.

4. The computer system of claim 2 , wherein the identified adjacency pairs are used to generate the sentence encoding representation.

5. The computer system of claim 1 , wherein the context-dependent vector and the context-independent vector are combined at least in part through a skip-gram model.

6. The computer system of claim 1 , wherein the context-independent database comprises mappings of various words to various context-independent vectors based upon definitions of the various words.

7. The computer system of claim 1 , wherein the text-based data is addressed to the particular user or generated by the particular user.

8. The computer system of claim 1 , wherein the text-based data comprises emails.

9. The computer system of claim 1 , wherein the executable instructions include instructions that are executable to configure the computer system to:

receive communication text input by the particular user;

map the communication text to the sentence encoding representation; and

based upon the sentence encoding representation, generate a collection of text that is associated with the sentence encoding representation, wherein the collection of text is responsive to the communication text.

10. The computer system of claim 9 , wherein the executable instructions include instructions that are executable to configure the computer system to:

identify a first part of a particular adjacency pair within the communication text; and

wherein the collection of text is based upon a second part of the particular adjacency pair.

11. A method for parsing bulk message information into intelligent question retrieval models, comprising:

receiving text-based data associated with a particular user;

parsing a word from the text-based data;

encoding the word into a context-dependent vector, wherein the context-dependent vector indicates the meaning of the word across a semantic space based upon the context of the word with the text-based data;

identifying within a context-independent database a context-independent vector that is associated with the word, wherein the context-independent database comprises mappings of various words to various context-independent vectors based upon definitions of the various words and wherein the context-independent vector is based upon definitions of the word and without insight into the text-based data;

generating an objective output by combining the context-dependent vector and the context-independent vector and a regularization function;

generating a sentence encoding representation by processing at least a portion of the text-based data through a high-level feature embedded convolutional semantic model to generate numerical representations of questions and answers within the text-based dataset, wherein the sentence encoding representation is generated at least in part based upon the objective output; and

storing the sentence encoding representation within a user-specific dataset that is associated with the particular user.

12. The method of claim 11 , further comprising:

parsing words from the text-based data; and

identifying adjacency pairs within the words.

13. The method of claim 12 , wherein identifying adjacency pairs comprises:

identifying the presence of a portion of the parsed words within a first entry in an adjacency pair database; and

identifying the presence of another portion of the parsed words within a second entry in the adjacency pair database, wherein the second entry to indicated as being complimentary to the first entry.

14. The method of claim 12 , wherein the identified adjacency pairs are used to generate the sentence encoding representation.

15. The method of claim 11 , wherein the context-dependent vector and the context-independent vector are combined at least in part through a skip-gram model.

16. The method of claim 11 , wherein the text-based data is addressed to the particular user or generated by the particular user.

17. The method of claim 11 , wherein the user-specific dataset consists of data related to the particular user.

18. The method of claim 11 , further comprising:

receiving communication text input by the particular user;

mapping the communication text to the sentence encoding representation; and

based upon the sentence encoding representation, generating a collection of text that is associated with the sentence encoding representation, wherein the collection of text is responsive to the communication text.

19. The computer system of claim 18 , wherein the executable instructions include instructions that are executable to configure the computer system to:

identifying a first part of a particular adjacency pair within the communication text; and

wherein the collection of text is based upon a second part of the particular adjacency pair.

20. A computer system for parsing bulk message information into intelligent question retrieval models, comprising:

one or more processors; and

one or more computer-readable media having stored thereon executable instructions that when executed by the one or more processors configure the computer system to perform at least the following:

receive text-based data associated with a particular user;

identify adjacency pairs within the words, wherein the adjacency pairs comprise two parts that are communicated by different parties in which a first part is conditionally related to a second part;

encode a word associated within the adjacency pair into a context-dependent vector, wherein the context-dependent vector indicates the meaning of the word across a semantic space based upon the context of the word with the text-based data;

identify within a context-independent database a context-independent vector that is associated with the word, wherein the context-independent database comprises mappings of various words to various context-independent vectors based upon definitions of the various words and wherein the context-independent vector is based upon definitions of the word and without insight into the text-based data; and

generate an objective output by combining the context-dependent vector and the context-independent vector and a regularization function.