Enterprise hypothesis orchestration

A computer implemented system for hypothesis orchestration, comprising:

• a user interface;
• one or more data repositories;
• a processor;
• and
• a non-transitory storage medium housing a plurality of instructions wherein the plurality of instructions when executed by the processor cause
• the processor to receive, via the user interface, a situation characterized by a set of linked inquiries,
• the processor to form a semantic graph pattern describing the set of linked inquiries for one or more pieces of evidence and thereafter search, using the semantic graph pattern, the one or more data repositories to identify the one or more pieces of evidence relevant to the set of linked inquiries wherein the one or more pieces of evidence are amassed by one or more collection resources,
• the processor to form a plurality of hypotheses based on evidence identified as relevant to the set of linked inquiries wherein each hypothesis is a proposed solution of the situation representing a unique path through the sematic graph pattern of linked pieces of evidence answering a subset of the set of linked inquiries,
• the processor to collect evidence identified as relevant to one or more of the plurality of hypotheses from the one or more data repositories forming, for each hypothesis, a set of collected pieces of evidence;
• the processor to evaluate each of the plurality of hypotheses with fuzzy logic using the set of collected pieces of evidence for that respective hypothesis to determine for each hypothesis a confidence value identifying a degree on which the set of collected pieces of evidence align with the situation, and
• responsive to the confidence value for one or more of the plurality of hypotheses failing to exceed a predefined threshold value,
• the processor to compare the set of collected pieces of evidence with the amassed one or more pieces of evidence relevant to the situation to identify, one or more missing pieces of evidence,
• the processor to isolate one or more common missing pieces of evidence from the one or more missing pieces of evidence that is relevant to two or more of the plurality of hypotheses that failed to exceed the predefined threshold value,
• the processor to form a value of information metric of missing information to the situation provided by each common missing piece of evidence,
• the processor to select at least one common missing piece of evidence from the one or more common missing pieces of evidence based on the respective value of information of the missing information to address the situation;
• the processor to initiate an information request to the one or more collection directing resources to seek information related to the situation based on the at least one common missing piece of evidence and on the value of information metric to reduce uncertainty associated with the plurality of hypotheses,
• the processor to receive the information related to the at least one common missing piece of evidence from the one or more collection resources;
• and
• the processor to refine the plurality of hypotheses relevant to the situation based on the received information related to the at least one common missing piece of evidence.

A non-transitory computer-readable storage medium tangibly embodying a program of instructions executable by a machine for hypothesis orchestration wherein said machine executes said program of instructions to perform operations of hypothesis orchestration, the operations comprising:

• receiving, by the machine via a user interface and rendered by the one or more processors, a situation characterized by a set of linked inquiries;
• searching, by the machine, one or more data repositories to identify one or more pieces of evidence relevant to the set of linked inquiries, wherein the one or more pieces of evidence are amassed by one or more collection resources and wherein searching includes forming a semantic graph describing the set of linked inquiries for the one or more pieces of evidence;
• forming, by the machine, a plurality of hypotheses based on evidence identified as relevant to the set of linked inquiries, wherein each hypothesis is a proposed solution of the situation representing a unique path through the sematic graph of linked pieces of evidence, answering a subset of the set of linked inquiries;
• collecting, by the machine, evidence identified as relevant to one or more of the plurality of hypotheses from the one or more data repositories forming, for each hypothesis, a set of collected pieces of evidence;
• evaluating, by the machine using fuzzy logic, each of the plurality of hypotheses using the set of collected pieces of evidence for that hypothesis thereby determining for each hypothesis a confidence value identifying a degree on which the set of collected pieces of evidence aligns with the situation;
• and
• responsive to the confidence value for one or more of the plurality of hypotheses failing to exceed a predefined threshold value,
• identifying, by the machine, one or more missing pieces of evidence by comparing the set of collected pieces of evidence with the amassed one or more pieces of evidence relevant to the situation,
• isolating, by the machine, one or more common missing pieces of evidence from the one or more missing pieces of evidence that is relevant to two or more of the plurality of hypotheses that failed to exceed the predefined threshold value,
• forming, by the machine, a value of information metric of missing information provided by each common missing piece of evidence to the situation,
• selecting, by the machine, at least one common missing piece of evidence from the one or more common missing pieces of evidence based on the respective value of information metric of the missing information to address the situation,
• initiating, by the machine, an information request to the one or more collection resources directing the one or more collection resources to seek information related to the situation based on the at least one common missing piece of evidence based and on the value of information metric to reduce uncertainty associated with one or more of the plurality of hypotheses,
• receiving, by the machine, the information related to the at least one common missing piece of evidence from the one or more collection resources;
• and
• refining, by the machine, the plurality of hypotheses relevant to the situation based on the received information related to the at least one common missing piece of evidence.