System and method for processing natural language statements

1. A computer implemented system for processing natural language statements, the system comprising:

a processor; and

a memory coupled to the processor storing processor readable instructions that, when executed, cause the processor to:

receive an entity index value associated with an identified entity and a time-scale associated with the entity index value;

provision a content prediction model defined by refined model parameters associated with the time-scale associated with the entity index value, wherein provisioning the content prediction model includes refining a prototypical topic transform and refining a prototypical entity transform based on the time-scale associated with the entity index value;

generate, based on the prototypical topic transform, an embedded content dimension vector based on an embedding data structure associated with statements made about the identified entity for biasing prototypical topics refined for the identified entity over the time-scale;

generate a content weight vector based on the embedding data structure and the prototypical entity transform for combining prototypical topics associated with one or more prototypical entities;

receive a statement history matrix including at least one content representation vector of an associated historical statement;

transform the statement history matrix with the generated content weight vector to provide a historical content dimension vector;

generate a content prediction score by generating a combination of the embedded content dimension vector and the historical content dimension vector and transforming the combination with an activation function; and

transmit a signal for communicating the content prediction score to update a machine learning model or identify novelty or impact of an identified statement received at the system.

2. The system of claim 1 , wherein the processor readable instructions, when executed, cause the processor to:

generate a predicted statement based on the content prediction score; and

determine a novelty score based on the identified statement received at the system and the predicted statement.

3. The system of claim 2 , wherein the processor readable instructions, when executed, cause the processor to:

combine a data representation of the identified statement with a set of historic statements associated with the statement history matrix to provide an updated statement history matrix;

generate an updated content prediction score and a subsequent predicted statement based on the updated statement history matrix; and

determine an impact score based on the predicted statement and the subsequent predicted statement.

4. The system of claim 3 , wherein the processor readable instructions, when executed, cause the processor to:

generate a signal for generating a communication based on the impact score.

5. The system of claim 4 , wherein the communication comprises at least one of a message highlighting the identified statement received at the system, a message excluding the identified statement received at the system, or a message associated with ranking the identified statement received at the system among a list of ranked statements.

6. The system of claim 1 , wherein generating the content prediction score includes:

receiving an average dimension vector associated with a global statement set; and

combining the average dimension vector with the embedded content dimension vector and the historical content dimension vector.

7. The system of claim 1 , wherein the at least one content representation vector is provided by at least one recurrent neural network.

8. The system of claim 7 , wherein the recurrent neural network is a gated recurrent unit.

9. The system of claim 1 , wherein the activation function includes a sigmoid function.

10. The system of claim 1 , wherein generating the embedded content dimension vector includes:

identifying a data subset of the embedding data structure associated with the entity index value to identify an entity embeddings vector; and

transforming the entity embeddings vector to the embedded content dimension vector associated with a weighted representation of content values.

11. The system of claim 1 , wherein generating the content weight vector based on the embedding data structure includes:

identifying a data subset of the embedding data structure associated with the entity index value to identify an entity embeddings vector; and

transforming the entity embeddings vector to the content weight vector to provide a condensed representation of the entity associated with the entity index value.

12. A computer implemented method for processing natural language statements, the method comprising:

receiving an entity index value associated with an identified entity and a time-scale associated with the entity index value;

provisioning a content prediction model defined by refined model parameters associated with the time-scale associated with the entity index value, wherein provisioning the content prediction model includes refining a prototypical topic transform and refining a prototypical entity transform based on the time-scale associated with the entity index value;

generating, based on the prototypical topic transform, an embedded content dimension vector based on an embedding data structure associated with statements made about the identified entity for biasing prototypical topics refined for the identified entity over the time-scale;

generating a content weight vector based on the embedding data structure and the prototypical entity transform for combining prototypical topics associated with one or more prototypical entities;

receiving a statement history matrix including at least one content representation vector of an associated historical statement;

transforming the statement history matrix with the generated content weight vector to provide a historical content dimension vector;

generating a content prediction score by generating a combination of the embedded content dimension vector and the historical content dimension vector and transforming the combination with an activation function; and

transmitting a signal for communicating the content prediction score to update a machine learning model or identify novelty or impact of an identified statement received at the system.

13. The computer implemented method of claim 12 , comprising:

generating a predicted statement based on the content prediction score; and

determining a novelty score based on the identified statement received at the system and the predicted statement.

14. The computer implemented method of claim 13 , comprising:

combining a data representation of the identified statement with a set of historic statements associated with the statement history matrix to provide an updated statement history matrix;

generating an updated content prediction score and a subsequent predicted statement based on the updated statement history matrix; and

determining an impact score based on the predicted statement and the subsequent predicted statement.

15. The computer implemented method of claim 14 , comprising:

generating a signal for generating a communication based on the impact score, wherein the communication comprises at least one of a message highlighting the identified statement received at the system, a message excluding the identified statement received at the system, or a message associated with ranking the identified statement received at the system among a list of ranked statements.

16. The computer implemented method of claim 12 , wherein generating the content prediction score includes:

receiving an average dimension vector associated with a global statement set; and

combining the average dimension vector with the embedded content dimension vector and the historical content dimension vector.

17. The computer implemented method of claim 12 , wherein the at least one content representation vector is provided by at least one recurrent neural network.

18. The computer implemented method of claim 17 , wherein the recurrent neural network is a gated recurrent network.

19. The computer implemented method of claim 12 , wherein the activation function includes a sigmoid function.

20. A non-transitory computer-readable medium or media having stored thereon machine interpretable instructions which, when executed by a processor, cause the processor to perform a computer implemented method for processing natural language statements, the method comprising:

receiving an entity index value associated with an identified entity and a time-scale associated with the entity index value;

provisioning a content prediction model defined by refined model parameters associated with the time-scale associated with the entity index value, wherein provisioning the content prediction model includes refining a prototypical topic transform and refining a prototypical entity transform based on the time-scale associated with the entity index value;

generating, based on the prototypical topic transform, an embedded content dimension vector based on an embedding data structure associated with statements made about the identified entity for biasing prototypical topics refined for the identified entity over the time-scale;

generating a content weight vector based on the embedding data structure and the prototypical entity transform for combining prototypical topics associated with one or more prototypical entities;

receiving a statement history matrix including at least one content representation vector of an associated historical statement;

transforming the statement history matrix with the generated content weight vector to provide a historical content dimension vector;

generating a content prediction score by generating a combination of the embedded content dimension vector and the historical content dimension vector and transforming the combination with an activation function; and

transmitting a signal for communicating the content prediction score to update a machine learning model or identify novelty or impact of an identified statement received at the system.