Method and system for discovering dynamic relations among entities

Abstract

Method, system, and programs for detecting dynamic relationship and discovering dynamic events. Data from a first data source is first received. At least one dynamic relation candidate is identified and each dynamic relation candidate involves multiple entities. The at least one dynamic relation candidate is identified based on temporal properties with respect to the entities exhibited in the data from the first data source. Dynamic relations are then extracted by corroborating the temporal properties of the entities involved in the at least one dynamic relation candidate with that of the same entities exhibited in data from a second data source. Then, a dynamic event that gives rise to the dynamic relations among different entities is detected.

1 . A method, implemented on a machine having at least one processor, storage, and a communication platform connected to a network for detecting a dynamic relationship and discovering a dynamic event, comprising the steps of:
 receiving, via the communication platform, data from a first data source;   identifying, by a dynamic relation candidate detector, at least one dynamic relation candidate, each of which involves a plurality of entities, based on temporal properties with respect to the one or more entities exhibited in the data from the first data source;   extracting, by a dynamic relation extractor, dynamic relations by corroborating the temporal properties of the entities involved in the at least one dynamic relation candidate with that of the same entities exhibited in data from a second data source;   detecting, by a dynamic event detector, a dynamic event that gives rise to the dynamic relations among different entities.   
     
     
         2 . The method of  claim 1 , wherein the first and second data sources are independent of each other. 
     
     
         3 . The method of  claim 1 , wherein the step of identifying comprises the steps of:
 constructing a first temporal profile for each of the entities identified in the data from the first source based on frequencies of occurrences of the entity in time in the data from the first data source;   identifying one or more peaks in a first temporal profile for each entity; and   extracting co-peaking entities, whose first temporal profiles exhibit peaks during a same time period with each other, as entities involved in a dynamic relation candidate.   
     
     
         4 . The method of  claim 1 , wherein the step of extracting comprises the steps of:
 constructing, for each entity involved in a dynamic relation candidate, a second temporal profile based on data from the second source based on frequencies of occurrences of the entity in time in the data from the second data source;   identifying one or more peaks in the second temporal profile for each entity;   identifying co-peaking entities, whose second temporal profiles exhibit peaks during a same time period;   confirming a dynamic relation candidate as a dynamic relation when co-occurrence of co-peaking temporal property is detected across data from both the first and the second data sources; and   generating a representation for each such confirmed dynamic relation.   
     
     
         5 . The method of  claim 1 , wherein the step of detecting comprises:
 obtaining a representation of a dynamic relation involving a plurality of entities and one or more connections, each of which linking two entities;   accessing one or more temporal constraints; and   identifying an event that gives rise to the dynamic relation based on the one or more temporal constraints and application thereof to the representation, wherein the event involves a set of entities and a set of connections linking the set of entities.   
     
     
         6 . The method of  claim 5 , wherein the one or more temporal constraints include a global temporal constraint and a local temporal constraint. 
     
     
         7 . The method of  claim 6 , wherein:
 the global temporal constraint requires that the set of connections be restricted within a pre-determined length of time; and   the local temporal constraint requires that the set of connections exhibits continuity in time and has one commonly connected entity.   
     
     
         8 . The method of  claim 1 , further comprising the step of generating, by an event/relation description generator, a representation of the dynamic event characterizing the dynamic event in terms of at least one of the entities involved in the dynamic event and the event itself. 
     
     
         9 . A system for detecting a dynamic relationship and discovering a dynamic event, comprising:
 a network communication platform connected to a network, through which data from a first data source can be obtained;   a dynamic relation candidate detector configured for identifying at least one dynamic relation candidate, each of which involves a plurality of entities, based on temporal properties with respect to the one or more entities exhibited in the data from the first data source;   a dynamic relation extractor configured for extracting dynamic relations by corroborating the temporal properties of the entities involved in the at least one dynamic relation candidate with that of the same entities exhibited in data from a second data source; and   a dynamic event detector configured for detecting a dynamic event that gives rise to the extracted dynamic relations.   
     
     
         10 . The system of  claim 9 , wherein the dynamic relation candidate detector comprises:
 a temporal profile generator configured for constructing a first temporal profile for each entity detected from data of the first data source based on frequencies of occurrences of the entity in time in the data from the first source;   a temporal peak identifier configured for identifying one or more peaks in a first temporal profile for each entity; and   a co-peaking detector configured for extracting co-peaking entities, whose first temporal profiles exhibit peaks during a same time period with each other, as entities involved in a dynamic relation candidate.   
     
     
         11 . The system of  claim 9 , wherein the dynamic relation extractor comprises:
 a temporal profile generator configured for constructing, for each entity involved in a dynamic relation candidate, a second temporal profile based on data from the second source based on frequencies of occurrences of the entity in time in the data from the second data source;   a temporal peak identifier configured for identifying one or more peaks in the second temporal profile for each entity;   a co-peaking detector configured for identifying co-peaking entities, whose second temporal profiles exhibit peaks during a same time period;   a co-occurrence corroboration mechanism configured for confirming a dynamic relation candidate as a dynamic relation when co-occurrence of co-peaking temporal property is detected across data from both the first and the second data sources; and   a dynamic relation representation generator configured for generating a representation for each such confirmed dynamic relation.   
     
     
         12 . The system of  claim 9 , wherein the step of detecting comprises:
 an event candidate detector configured for obtaining a representation of dynamic relations involving a plurality of entities and one or more connections, each of which linking two entities;   one or more event consolidation units, each of which is configured for enforcing a temporal constraint on the dynamic relation to detect a candidate dynamic event that satisfies the temporal constraint; and   an event identification controller configured for identifying a dynamic event that gives rise to the dynamic relations based on corresponding candidate dynamic event from each event consolidation unit, wherein the dynamic event involves a set of entities and a set of connections linking the set of entities.   
     
     
         13 . The system of  claim 9 , further comprising an event/relation description generator configured for generating a representation of the dynamic event characterizing the dynamic event in terms of at least one of the entities involved in the dynamic event and the event itself. 
     
     
         14 . A machine readable non-transitory and tangible medium having information recorded thereon for detecting a dynamic relationship and discovering a dynamic event, wherein the information, when read by the machine, causes the machine to perform the following:
 receiving data from a first data source;   identifying at least one dynamic relation candidate, each of which involves a plurality of entities, based on temporal properties with respect to the one or more entities exhibited in the data from the first data source;   extracting dynamic relations by corroborating the temporal properties of the entities involved in the at least one dynamic relation candidate with that of the same entities exhibited in data from a second data source;   detecting a dynamic event that gives rise to the dynamic relations among different entities.   
     
     
         15 . The medium of  claim 14 , wherein the first and second data sources are independent of each other. 
     
     
         16 . The medium of  claim 14 , wherein the step of identifying comprises the steps of:
 constructing a first temporal profile for each of the entities identified in the data from the first source based on frequencies of occurrences of the entity in time in the data from the first data source;   identifying one or more peaks in a first temporal profile for each entity; and   extracting co-peaking entities, whose first temporal profiles exhibit peaks during a same time period with each other, as entities involved in a dynamic relation candidate.   
     
     
         17 . The medium of  claim 14 , wherein the step of extracting comprises the steps of:
 constructing, for each entity involved in a dynamic relation candidate, a second temporal profile based on data from the second source based on frequencies of occurrences of the entity in time in the data from the second data source;   identifying one or more peaks in the second temporal profile for each entity;   identifying co-peaking entities, whose second temporal profiles exhibit peaks during a same time period;   confirming a dynamic relation candidate as a dynamic relation when co-occurrence of co-peaking temporal property is detected across data from both the first and the second data sources; and   generating a representation for each such confirmed dynamic relation.   
     
     
         18 . The medium of  claim 14 , wherein the step of detecting comprises:
 obtaining a representation of a dynamic relation involving a plurality of entities and one or more connections, each of which linking two entities;   accessing one or more temporal constraints; and   identifying an event that gives rise to the dynamic relation based on the one or more temporal constraints and application thereof to the representation, wherein the event involves a set of entities and a set of connections linking the set of entities.   
     
     
         19 . The medium of  claim 18 , wherein the one or more temporal constraints include a global temporal constraint and a local temporal constraint, wherein:
 the global temporal constraint requires that the set of connections be restricted within a pre-determined length of time; and   the local temporal constraint requires that the set of connections exhibits continuity in time and has one commonly connected entity.   
     
     
         20 . The medium of  claim 14 , wherein the information, when read by the machine, further causes the machine to perform the step of generating a representation of the dynamic event characterizing the dynamic event in terms of at least one of the entities involved in the dynamic event and the event itself.