Exercising artificial intelligence by refining model output
The improved exercise of artificial intelligence. Raw output data is obtained by applying an input data set to an artificial intelligence (AI). Such raw output data is sometimes difficult to interpret. The principles defined herein provide a systematic way to refine the output for a wide variety of AI models. An AI model collection characterization structure is utilized for purpose of refining AI model output so as to be more useful. The characterization structure represents, for each of multiple and perhaps numerous AI models, a refinement of output data that resulted from application of an AI model to input data. Upon obtaining output data from the AI model, the appropriate refinement may then be applied. The refined data may then be semantically indexed to provide a semantic index. The characterization structure may also provide tailored information to allow for intuitive querying against the semantic index.
1. A computing system comprising:
one or more processors; and
one or more computer-readable hardware storage devices storing computer-executable instructions that are executable by the one or more processors to cause the computing system to at least:
obtain results of an input data set applied to an artificial intelligence (AI) model;
access at least a portion of an AI characterization structure, the AI characterization structure comprising:
a plurality of operational AI model representations, wherein an operational AI model representation is associated with a combination of an input data set type included in a plurality of input data set types and an AI model type included in a plurality of AI model types, the input data set type being a type of data identified for the input data set that was applied to the AI model, the AI model type being a type of model identified for the AI model such that a subsequently selected refinement definition is selected based on the combination of the input data set type and the AI model type that were identified for the input data set and the AI model, and
a plurality of refinement definitions, each refinement definition being associated with at least one operational AI model representation included in the plurality of AI model representations;
determine the input data set type of the input data set and the AI model type of the AI model;
identify the operational AI model representation included in the plurality of operational AI model representations, wherein the one operational AI model representation is identified based on a determination that the one operational AI model representation is applicable to the obtained results based on a combination of the input data set type and the AI model type associated with the obtained results;
select a refinement definition included in the plurality of refinement definitions based on the identified one operational AI model representation, which is identified as a result of the obtained results being associated with the combination of the input data set type and the AI model type; and
use the selected refinement definition to refine the obtained results, wherein refining the obtained results using the selected refinement definition is further augmented by machine learning analysis to determine a granularity at which to refine the obtained results.
2. The computing system in accordance with claim 1 , wherein execution of the computer-executable instructions further causes the computing system to at least:
semantically index the refined results to generate a semantic index, the artificial intelligence (AI) characterization structure further comprising a set of one or more operators and/or terms that a query engine may use to query against the semantic index; and
cause at least a portion of the set of one or more operators and/or terms to be communicated to the query engine.
3. The computing system in accordance with claim 1 , wherein execution of computer-executable instructions further causes the computing system to at least:
semantically index the refined results to generate a semantic index, the artificial intelligence (AI) characterization structure further comprising a set of one or more visualizations that a visualization engine may use to visualize to a user responses to queries against the semantic index; and
cause at least a portion of the set of one or more visualizations to be communicated to the visualization engine.
4. A method for a computing system to exercise artificial intelligence (AI), the method comprising:
obtaining results of an input data set applied to an artificial intelligence (AI) model;
accessing at least a portion of an AI characterization structure, the AI characterization structure comprising:
a plurality of operational AI model representations, wherein an operational AI model representation is associated with a combination of an input data set type included in a plurality of input data set types and an AI model type included in a plurality of AI model types, the input data set type being a type of data identified for the input data set that was applied to the AI model, the AI model type being a type of model identified for the AI model such that a subsequently selected refinement definition is selected based on the combination of the input data set type and the AI model type that were identified for the input data set and the AI model, and
a plurality of refinement definitions, each refinement definition being associated with at least one operational AI model representation included in the plurality of AI model representations;
determining the input data set type of the input data set and the AI model type of the AI model;
identifying the operational AI model representation included in the plurality of operational AI model representations, wherein the one operational AI model representation is identified based on a determination that the one operational AI model representation is applicable to the obtained results based on a combination of the input data set type and the AI model type associated with the obtained results;
selecting a refinement definition included in the plurality of refinement definitions based on the identified one operational AI model representation, which is identified as a result of the obtained results being associated with the combination of the input data set type and the AI model type; and
using the selected refinement definition to refine the obtained results, wherein refining the obtained results using the selected refinement definition is further augmented by machine learning analysis to determine a granularity at which to refine the obtained results.
5. The method in accordance with claim 4 , further comprising:
semantically indexing the refined results to generate a semantic index.
6. The method in accordance with claim 5 , further comprising:
using the semantic index to present a suggested query to a user.
7. The method in accordance with claim 5 , the artificial intelligence (AI) characterization structure further comprising a set of one or more operators that a query engine may use to query against the semantic index, the method further comprising:
causing at least a portion of the set of one or more operators to be communicated to the query engine.
8. The method in accordance with claim 5 , the artificial intelligence (AI) characterization structure further comprising a set of one or more terms that a query engine may use to query against the semantic index, the method further comprising:
causing at least a portion of the set of one or more terms to be communicated to the query engine.
9. The method in accordance with claim 5 , the artificial intelligence (AI) characterization structure further comprising a set of one or more visualizations that a visualization engine may use to visualize to a user responses to queries against the semantic index, the method further comprising:
causing at least a portion of the set of one or more visualizations to be communicated to the visualization engine.
10. The method in accordance with claim 4 , the artificial intelligence (AI) characterization structure further representing, for each of a plurality of AI model and input data set type combinations, a refinement of results of data applied to an AI model, wherein the refining the obtained results is at least based on the refinement represented in the characterization structure for the combination of the AI model and the input data set.
11. The method in accordance with claim 4 , the refining of the obtained results also based on hints specific to the AI model.
12. The method in accordance with claim 11 , the hints specific to the AI model being within a model-specific data structure that is associated with the AI model.
13. The method in accordance with claim 4 , the refining of the obtained results also based on machine learning analysis of prior refinements of obtained results of data applied to an AI model.
14. The method in accordance with claim 4 , the refining of the obtained results also based on machine learning analysis of prior refinements of obtained results of the input data set applied to an AI model.
15. The method in accordance with claim 4 , the refining of the obtained results also based on machine learning analysis of prior refinements of obtained results of data applied to an AI model when those obtained results are provided for a particular user, such that the refining is specific to the particular user.
16. The method in accordance with claim 4 , the AI model comprising a machine learning model.
17. The method in accordance with claim 4 , the obtained results being first obtained results, the input data set being first input data set of a first data set type, the obtained results being first obtained results, the method further comprising:
obtaining results of a second input set data of a second data set type applied to the AI model to obtain second obtained results; and
refining the second obtained results at least based on the refinement represented in the AI characterization structure for the AI model.
18. The method in accordance with claim 4 , the obtained results being first obtained results, the input data set being first input data set of a first data set type, the refinement represented in the characterization structure for the AI model being a first refinement that is applicable for the AI model and input data set of the first data set type, the obtained results being first obtained results, the method further comprising:
obtaining results of a second input data set of a second data set type applied to the AI model to obtain second obtained results; and
refining the second obtained results at least based on a second refinement represented in the AI characterization structure for the AI model and input data set of the second data set type, the second refinement being different than the first refinement.
19. The method in accordance with claim 4 , the AI model being a first AI model, the method further comprising:
obtaining results of a second input data set applied to a second AI model, the second AI model also being one of the plurality of AI models; and
refining the obtained results from the second AI model at least based on a refinement represented in an AI characterization structure for the second AI model.
20. A computer program product comprising one or more computer-readable storage media having thereon computer-executable instructions that are executable by one or more processors of a computing system to cause the computing system to exercise artificial intelligence (AI) by causing the computing system to at least:
obtain results of an input data set applied to an artificial intelligence (AI) model;
access at least a portion of an AI characterization structure, the AI characterization structure comprising:
a plurality of operational AI model representations, wherein an operational AI model representation is associated with a combination of an input data set type included in a plurality of input data set types and an AI model type included in a plurality of AI model types, the input data set type being a type of data identified for the input data set that was applied to the AI model, the AI model type being a type of model identified for the AI model such that a subsequently selected refinement definition is selected based on the combination of the input data set type and the AI model type that were identified for the input data set and the AI model, and
a plurality of refinement definitions, each refinement definition being associated with at least one operational AI model representation included in the plurality of AI model representations;
determine the input data set type of the input data set and the AI model type of the AI model;
identify the operational AI model representation included in the plurality of operational AI model representations, wherein the one operational AI model representation is identified based on a determination that the one operational AI model representation is applicable to the obtained results based on a combination of the input data set type and the AI model type associated with the obtained results;
select a refinement definition included in the plurality of refinement definitions based on the identified one operational AI model representation, which is identified as a result of the obtained results being associated with the combination of the input data set type and the AI model type; and
use the selected refinement definition to refine the obtained results, wherein refining the obtained results using the selected refinement definition is further augmented by machine learning analysis to determine a granularity at which to refine the obtained results.