Provenance analysis systems and methods

Abstract

Provenance analysis systems and methods. Datums representing relationships between entities can be stored in a knowledge store. Datums can be received from agents as agents perform activities. Activity records are be stored in a provenance graph, the activity record and associate received datums with any input datums used in the activity. Provenance subgraphs can be retrieved by traversing the provenance graph for selected datums and presented through a user interface. Provenance subgraphs can be augmented with trust modifiers determined based on attributions, confidences, and refutations provided by a user. Trust modifiers can be propagated downstream to enable the addressing of junctions in variable confidence.

What is claimed is: 
     
         1 . A provenance analysis system comprising:
 a knowledge store comprising a plurality of datum records, each datum record comprising a datum, the datum comprising an indicator of a relationship between a first entity and a second entity;   an agent interface configured to:
 receive a datum from an agent; 
 store an activity record in a provenance graph, the activity record comprising an association with the received datum and any input datums used by the agent to generate the received datum; 
   a provenance subgraph retriever configured to:
 receive a selected datum, and 
 retrieve a provenance subgraph for the selected datum by recursively identifying each activity record in the provenance graph associated with the selected datum and each activity record in the provenance subgraph of any input datum of an activity record associated with the selected datum; and 
   a provenance visualizer comprising a display and a user input interface for receiving a selection of a datum, the provenance visualizer configured to display the retrieved provenance graph.   
     
     
         2 . The system of  claim 1 , wherein the agent interface is further configured to:
 search the knowledge store to determine if a datum record comprising the received datum exists, and in response to determining that a datum record comprising the received datum does not exist, storing a datum record comprising the received datum in the knowledge store.   
     
     
         3 . The system of  claim 1 , further comprising:
 a confidence propagator configured to:
 receive a trust parameter comprising at least one activity attribute, and 
 apply the trust parameter to the provenance subgraph by augmenting each applicable activity record of the provenance subgraph with a trust modifier based on the trust parameter; and 
   wherein an activity record of the provenance subgraph is an applicable activity record if the activity record has an activity attribute that matches an activity attribute of the trust parameter or if the activity record has an input datum that is associated with an applicable activity record.   
     
     
         4 . The system of  claim 3 , wherein augmenting each applicable activity record of the provenance subgraph with a trust modifier based on the trust parameter comprises:
 for each activity record found in a search of the provenance subgraph for any activity records having an activity attribute that matches an activity attribute of the trust parameter:
 calculating a trust modifier value based on the trust parameter and any existing trust modifier value associated with the activity record; 
 and iteratively applying the trust modifier value to each child activity record having an input datum that is associated with the activity record. 
   
     
     
         5 . The system of  claim 4 , wherein the trust parameter comprises a confidence value and calculating a trust modifier value based on the trust parameter and any existing trust modifier value associated with the activity record comprises:
 associating a trust modifier value of 1 the activity record if no existing trust modifier value is associated with the activity record; and   multiplying the trust modifier value by each trust modifier value associated with the activity record.   
     
     
         6 . The system of  claim 3 , wherein the user input interface is further configured to receive the trust parameter, and the provenance visualizer is configured to alter the display of the provenance subgraph based on the trust modifier of each activity record. 
     
     
         7 . The system of  claim 3 , wherein the trust parameter indicates a confidence in a datum. 
     
     
         8 . The system of  claim 3 , wherein each activity record further comprises an identifier of the agent and the trust parameter comprises a value indicating a confidence in activities performed by an agent. 
     
     
         9 . The system of  claim 3 , wherein each activity record further comprises a class of the agent and the trust parameter comprises a value indicating a confidence in activities performed a class of agents. 
     
     
         10 . The system of  claim 3 , wherein each activity record further comprises a class of operation performed by the agent and the trust parameter comprises a value indicating a confidence in a class of operations. 
     
     
         11 . The system of  claim 3 , wherein each activity record further comprises a source of any input datum associated with an activity, and the trust parameter comprises a value indicating a confidence in a source. 
     
     
         12 . The system of  claim 1  further comprising a conflict detector, the conflict detector configured to indicate one or more conflicting datums in the provenance graph. 
     
     
         13 . A method for visualizing the provenance of a datum record in a knowledge store, the method comprising:
 providing a knowledge store for storage of a plurality of datum records, each datum record comprising a unique datum, the datum comprising an indicator of a relationship between a first entity and a second entity;   receiving a datum from an agent;
 storing an activity record in a provenance graph, the activity record comprising an association with the received datum and any input datums used by the agent to generate the received datum; 
   receiving an identifier of a selected datum;
 retrieving a provenance subgraph for the selected datum by recursively identifying each activity record in the provenance graph associated with the selected datum and each activity record in the provenance graph of any input datum of an activity record associated with the selected datum; and 
   displaying the retrieved provenance subgraph, by a provenance visualizer comprising a display and a user input interface for receiving a selection of a datum record.   
     
     
         14 . The method of  claim 13 , further comprising
 searching the knowledge store to determine if a datum record comprising the received datum exists, and in response to determining that a datum record comprising the received datum does not exist, storing a datum record comprising the received datum in the knowledge store.   
     
     
         15 . The method of  claim 13 , further comprising:
 receiving a trust parameter comprising a t least one activity attribute; and   applying the trust parameter to the provenance subgraph by augmenting each applicable activity record of the provenance subgraph with a trust modifier based on the trust parameter; and   wherein an activity record of the provenance subgraph is an applicable activity record if the activity record has an activity attribute that matches an activity attribute of the trust parameter or if the activity record has an input datum that is associated with an applicable activity record.   
     
     
         16 . The method of  claim 15 , wherein augmenting each applicable activity record of the provenance subgraph with a trust modifier based on the trust parameter comprises:
 for each activity record found in a search of the provenance subgraph for any activity records having an activity attribute that matches an activity attribute of the trust parameter:
 calculating a trust modifier value based on the trust parameter and any existing trust modifier value associated with the activity record; 
 and iteratively applying the trust modifier to each child activity record having an input datum that is associated with the activity record. 
   
     
     
         17 . The method of  claim 15 , wherein the trust parameter comprises a confidence value and calculating a trust modifier value based on the trust parameter and any existing trust modifier value associated with the activity record comprises:
 associating a trust modifier value of 1 the activity record if no existing trust modifier value is associated with the activity record; and   multiplying the confidence value by each trust modifier value associated with the activity record.   
     
     
         18 . The method of  claim 15 , wherein the trust parameter is received via the user input interface and further comprising altering the display of the provenance subgraph based on the trust modifier of each activity record. 
     
     
         19 . The method of  claim 15 , wherein the trust parameter indicates a confidence in a datum. 
     
     
         20 . The method of  claim 15 , wherein each activity record further comprises an identifier of the agent and the trust parameter comprises a value indicating a confidence in activities performed by an agent. 
     
     
         21 . The method of  claim 15 , wherein each activity record further comprises a class of the agent and the trust parameter comprises a value indicating a confidence in activities performed a class of agents. 
     
     
         22 . The method of  claim 15 , wherein each activity record further comprises a class of operation performed by the agent and the trust parameter comprises a value indicating a confidence in a class of operations. 
     
     
         23 . The method of  claim 15 , wherein each activity record further comprises a source of any input datum, and the trust parameter comprises a value indicating a confidence in a source. 
     
     
         24 . The method of  claim 13 , further comprising a detecting one or more conflicting datum records, and altering the display to indicate the one or more conflicting datum records.