Accelerated reasoning graph evaluation
Embodiments disclosed herein relate to methods, systems, and computer programs for automatically determining an outcome associated with a reasoning graph, based on one or more data sets. The methods, systems, and computer programs compare hash values associated with different data sets to determine if they match to assign the outcome associated with a pre-existing hash to the later provided hash and data set associated therewith.
1. A method comprising:
upon receiving an input data set for a reasoning graph, obtaining, by a processor of an automated reasoning system configured for evaluating data against reasoning graphs based on following logical paths defined by the reasoning graphs, a plurality of hashes for the input data set, wherein each hash is generated from a subset of the input data set;
identifying, by the processor of the automated reasoning system, a particular correlation that is stored in a database that is associated with the reasoning graph, wherein the database stores a plurality of correlations between known hashes and identifiers for portions of the reasoning graph, and wherein the particular correlation correlates a particular hash for the input data set with a discrete decision identifier for a discrete decision within the reasoning graph; and
performing, by the processor of the automated reasoning system, a graph evaluation of the reasoning graph with the input data set, wherein the graph evaluation starts from the discrete decision identified by the discrete decision identifier and reaches an insight at an end of the reasoning graph,
wherein the graph evaluation from the discrete decision to the insight is associated with less computational resource consumption and time compared to a different graph evaluation with the input data set from a beginning function of the reasoning graph to the insight, due to avoiding evaluation of at least one reasoning function between the beginning function and the discrete decision.
2. The method of claim 1 , further comprising:
communicating the insight reached by the graph evaluation to a requesting entity associated with the input data set.
3. The method of claim 1 , wherein the portions of the reasoning graph include discrete decisions of the reasoning graph, leaf nodes of the reasoning graph, and reasoning paths through the reasoning graph.
4. The method of claim 1 , further comprising:
determining that a second hash for a second input data set fails to match any of the known hashes in the database; and
performing a second graph evaluation of the reasoning graph with the second input data set, wherein the second graph evaluation starts from the beginning function of the reasoning graph.
5. The method of claim 4 , further comprising:
storing a new correlation in the database, wherein the new correlation correlates the second hash with a new outcome that results from the second graph evaluation.
6. The method of claim 1 , wherein obtaining the plurality of hashes includes generating the plurality of hashes using a hash function locally implemented at the automated reasoning system.
7. The method of claim 1 , wherein the graph evaluation is performed using other subsets of the input data set that are not included in the subset from which the particular hash is generated.
8. A system comprising:
at least one processor; and
a memory storing executable instructions that, when executed by the at least one processor, cause the system to:
obtain a plurality of hashes for an input data set for a reasoning graph, wherein each hash is generated from a subset of the input data set;
reference a database that is associated with the reasoning graph to identify a particular correlation, wherein the database stores a plurality of correlations between known hashes and identifiers for portions of the reasoning graph, and wherein the particular correlation correlates a particular hash for the input data set with a discrete decision identifier for a discrete decision of the reasoning graph; and
cause a graph evaluation of the reasoning graph with the input data set, wherein the graph evaluation begins from the discrete decision identified by the discrete decision identifier and reaches an insight at an end of the reasoning graph,
wherein the graph evaluation from the discrete decision to the insight is associated with less computational resource consumption and time compared to a different graph evaluation with the input data set from a beginning function of the reasoning graph to the insight, due to avoiding evaluation of at least one reasoning function between the beginning function and the discrete decision.
9. The system of claim 8 , wherein the executable instructions further cause the system to communicate the insight reached by the graph evaluation to a requesting entity associated with the input data set.
10. The system of claim 8 , wherein the portions of the reasoning graph include discrete decisions of the reasoning graph, leaf nodes of the reasoning graph, and reasoning paths through the reasoning graph.
11. The system of claim 8 , wherein the executable instructions further cause the system to:
based on failing to identify a correlation with a hash that matches the hashes for the input data set, cause the graph evaluation to begin from the beginning function of the reasoning graph.
12. The system of claim 11 , wherein the executable instructions further cause the system to:
store, in the database, a new correlation that correlates an outcome with a hash for the input data set, the outcome resulting from the graph evaluation that began from the beginning function.
13. The system of claim 8 , wherein obtaining the plurality of hashes includes generating the plurality of hashes using a hash function locally implemented at the system.
14. The system of claim 8 , wherein causing the graph evaluation includes performing the graph evaluation using other subsets of the input data set that are not included in the subset from which the particular hash is generated.
15. A non-transitory computer program storage medium storing instructions that, when executed by a processor, cause the processor to:
obtain a plurality of hashes for an input data set for a reasoning graph, wherein each hash is generated from a subset of the input data set;
identify a discrete decision of the reasoning graph that is correlated with a particular hash for the input data set, wherein the discrete decision is identified based on a database storing a correlation between a discrete decision identifier for the discrete decision and a known hash that matches the particular hash; and
perform a graph evaluation of the reasoning graph with the input data set, the graph evaluation starting from the discrete decision and reaching an insight at an end of the reasoning graph,
wherein the graph evaluation from the discrete decision to the insight is associated with less computational resource consumption and time compared to a different graph evaluation with the input data set from a beginning function of the reasoning graph to the insight, due to avoiding evaluation of at least one reasoning function between the beginning function and the discrete decision.
16. The non-transitory computer program storage medium of claim 15 , wherein the instructions further cause the processor to:
communicate the insight reached by the graph evaluation to a requesting entity associated with the input data set.
17. The non-transitory computer program storage medium of claim 15 , wherein the database stores a plurality of correlations between known hashes and portions of the reasoning graph, wherein the portions include discrete decisions, leaf nodes, and reasoning paths.
18. The non-transitory computer program storage medium of claim 15 , wherein the instructions further cause the processor to:
based on failing to identify a correlation with a hash that matches the hashes for the input data set, cause the graph evaluation to begin from the beginning function of the reasoning graph.
19. The non-transitory computer program storage medium of claim 18 , wherein the instructions further cause the processor to:
store, in the database, a new correlation that correlates an outcome with a hash for the input data set, the outcome resulting from the graph evaluation that began from the beginning function.
20. The non-transitory computer program storage medium of claim 15 , wherein obtaining the plurality of hashes includes generating the plurality of hashes using a hash function.