Automatically generating reasoning graphs

1. A computing system comprising at least one processor and a memory storing instructions that, when executed by the at least one processor, cause the computing system to:

identify a reasoning graph that includes a plurality of nodes that are defined in a first language format and that are arranged along one or more reasoning paths in the reasoning graph, wherein each node describes a first reasoning function according to the first language format;

determine, for a given node of the plurality of nodes, a second function in a second language format to substitute for the first reasoning function of the given node using a transliteration library that defines one-to-one correlations between a first portfolio of reasoning functions in the first language format and a second portfolio of reasoning functions in the second language format;

generate a representation of the reasoning graph, the representation being defined in the second language format, wherein the representation includes at least one node that describes the second function that substitutes a particular first reasoning function described by a corresponding node of the reasoning graph, and wherein generating the representation of the reasoning graph comprises path-wise progressing through each of the one or more reasoning paths to sequentially substitute the first reasoning function of each node along each of the one or more reasoning paths; and

output the representation of the reasoning graph.

2. The computing system of claim 1 , wherein the representation of the reasoning graph retains a structure of the reasoning graph with respect to reasoning paths that connect the plurality of nodes in the reasoning graph and relative positions of the plurality of nodes within the reasoning graph.

3. The computing system of claim 1 , wherein the first language format is a human readable text format, and the first reasoning function of each node is defined in plain language.

4. The computing system of claim 1 , wherein the second language format is a machine readable code format, and the representation of the reasoning graph includes machine readable executable code that implements the reasoning graph.

5. The computing system of claim 4 , wherein the representation of the reasoning graph is generated based on generating an intermediate representation of the reasoning graph according to a templating language format and generating the representation of the reasoning graph from the intermediate representation.

6. The computing system of claim 5 , wherein a first compiler is used to generate the intermediate representation and a second compiler is used to generate the representation of the reasoning graph.

7. The computing system of claim 1 , wherein the representation of the reasoning graph includes one or more external links to a database that stores data used to evaluate the second function in the representation of the reasoning graph.

8. The computing system of claim 1 , further comprising determining an insight based on traversing the representation of the reasoning graph until reaching a leaf node of the representation of the reasoning graph that encodes the insight.

9. A method comprising:

identifying, by a processor, a reasoning graph that includes a plurality of nodes that are defined in a first language format and that are arranged along one or more reasoning paths in the reasoning graph, wherein each node describes a first reasoning function according to the first language format;

determining, by the processor, for a given node of the plurality of nodes, a second function in a second language format to substitute for the first reasoning function of the given node using a transliteration library that defines one-to-one correlations between a first portfolio of reasoning functions in the first language format and a second portfolio of reasoning functions in the second language format;

generating, by the processor, a representation of the reasoning graph, the representation being defined in the second language format, wherein the representation includes at least one node that describes the second function that substitutes a particular first reasoning function described by a corresponding node of the reasoning graph, and wherein generating the representation of the reasoning graph comprises path-wise progressing through each of the one or more reasoning paths to sequentially substitute the first reasoning function of each node along each of the one or more reasoning paths; and

outputting, by the processor, the representation of the reasoning graph.

10. The method of claim 9 , wherein the representation of the reasoning graph retains a structure of the reasoning graph with respect to reasoning paths that connect the plurality of nodes in the reasoning graph and relative positions of the plurality of nodes within the reasoning graph.

11. The method of claim 9 , wherein the first language format is a human readable text format, and the first reasoning function of each node is defined in plain language.

12. The method of claim 9 , wherein the second language format is a machine readable code format, and the representation of the reasoning graph includes machine readable executable code that implements the reasoning graph.

13. The method of claim 12 , wherein the representation of the reasoning graph is generated based on generating an intermediate representation of the reasoning graph according to a templating language and generating the representation of the reasoning graph from the intermediate representation.

14. The method of claim 13 , wherein a first compiler is used to generate the intermediate representation and a second compiler is used to generate the representation of the reasoning graph.

15. The method of claim 9 , wherein the representation of the reasoning graph includes one or more external links to a database that stores data used to evaluate the second function in the representation of the reasoning graph.

16. The method of claim 9 , further comprising determining an insight based on traversing the representation of the reasoning graph until reaching a leaf node of the representation of the reasoning graph that encodes the insight.

17. At least one non-transitory computer readable medium storing executable instructions configured to cause a processor to:

identify a reasoning graph that includes a plurality of nodes that are defined in a first language format and that are arranged along one or more reasoning paths in the reasoning graph, wherein each node describes a first reasoning function according to the first language format;

determine, for a given node of the plurality of nodes, a second function in a second language format to substitute for the first reasoning function of the given node using a transliteration library that defines one-to-one correlations between a first portfolio of reasoning functions in the first language format and a second portfolio of reasoning functions in the second language format;

generate a representation of the reasoning graph, the representation being defined in the second language format, wherein the representation includes at least one node that describes the second function that substitutes a particular first reasoning function described by a corresponding node of the reasoning graph, and wherein generating the representation of the reasoning graph comprises path-wise progressing through each of the one or more reasoning paths to sequentially substitute the first reasoning function of each node along each of the one or more reasoning paths; and

output the representation of the reasoning graph.

18. The at least one non-transitory computer readable medium of claim 17 , wherein the second language format is a machine readable code format, and the representation of the reasoning graph includes machine readable executable code that implements the reasoning graph.

19. The at least one non-transitory computer readable medium of claim 18 , wherein the representation of the reasoning graph is generated based on generating an intermediate representation of the reasoning graph according to a templating language and generating the representation of the reasoning graph from the intermediate representation.

20. The at least one non-transitory computer readable medium of claim 19 , wherein a first compiler is used to generate the intermediate representation and a second compiler is used to generate the representation of the reasoning graph.