System for inferring network dynamics and sources within the network

Patent metadata

Field	Value
Publication number	US-10652104-B1
Application number	US-201715782668-A
Country	US
Kind code	B1
Filing date	Oct 12, 2017
Priority date	Oct 12, 2016
Publication date	May 12, 2020
Grant date	May 12, 2020

Abstract

Described is a system for inferring network dynamics and their sources within the network. During operation, a vector representation is generated based on states of agents in a network. The vector representation including attribute vectors that correspond to the states of the agents in the network. A matrix representation is then generated based on the changing states of agents by packing the attribute vectors at each time step into an attribute matrix. Time-evolving states of the agents are learned using dictionary learning. Influential source agents in the network are then identified by performing dimensionality reduction on the attribute matrix. Finally, in some aspects, an action is executed based on the identity of the influential source agents. For example, marketing material may be directed to a source agent's online account, or the source agent's online account can be deactivated or terminated or some other desired action can be taken.

First claim

What is claimed is:

1. A system for inferring network dynamics and their sources within a network, the system comprising:
one or more processors and a memory, the memory being a non-transitory computer-readable medium having executable instructions encoded thereon, such that upon execution of the instructions, the one or more processors perform operations of:
generating a vector representation based on states of agents in a network, the vector representation including attribute vectors that correspond to the states of the agents in the network;
generating an attribute matrix based on the changing states of agents by packing the attribute vectors at each time step into the attribute matrix;
learning time-evolving states of agents using dictionary learning, wherein learning time-evolving states of agents using dictionary learning is performed through non-negative matrix factorization (NMF);
identifying influential source agents in the network by performing dimensionality reduction on the attribute matrix; and
executing an action based on the identity of the influential source agents.

2. The system as set forth in claim 1 , wherein in generating a vector representation based on the states of agents in a network, the network includes nodes and edges connecting the nodes, the nodes representing agents in the network and the edges representing actions by the agents in the network.

3. The system as set forth in claim 1 , wherein if the attributes at a given node are a time series, attribute vectors at each time step can be packed into a matrix of attributes X that is N by t dimensional, where t is the number of time steps in a signal and N is a number of nodes in the network.

4. The system as set forth in claim 1 , wherein the changing states of agents is based on a time-series of each agent.

5. The system as set forth in claim 1 , wherein learning the time-evolving states of agents using dictionary learning is performed using the attribute matrix.

6. The system as set forth in claim 1 , wherein learning the time-evolving states of agents using dictionary learning is performed in parallel using submatrices of the attribute matrix.

7. The system as set forth in claim 1 , further comprising an operation of acquiring data from a social network server that includes states of agents in a network.

8. The system as set forth in claim 1 , wherein executing an action further comprises an operation of directing marketing information to be displayed on a social media account associated with the influential source agents.

9. The system as set forth in claim 1 , wherein executing an action further comprises an operation of providing the identity of the influential source agents via a display.

10. The system as set forth in claim 1 , wherein executing an action further comprises an operation of deactivating a social media account associated with the influential source agents.

11. The system as set forth in claim 1 , wherein the vector representation is generated by converting a graph representation that includes agents, connections between agents, and states of agents.

12. A computer program product for inferring network dynamics and their sources within a network, the computer program product comprising:
a non-transitory computer-readable medium having executable instructions encoded thereon, such that upon execution of the instructions by one or more processors, the one or more processors perform operations of:
generating a vector representation based on states of agents in a network, the vector representation including attribute vectors that correspond to the states of the agents in the network;
generating an attribute matrix based on the changing states of agents by packing the attribute vectors at each time step into the attribute matrix;
learning time-evolving states of agents using dictionary learning, wherein learning time-evolving states of agents using dictionary learning is performed through non-negative matrix factorization (NMF);
identifying influential source agents in the network by performing dimensionality reduction on the attribute matrix; and
executing an action based on the identity of the influential source agents.

13. The computer program product as set forth in claim 12 , wherein in generating a vector representation based on the states of agents in a network, the network includes nodes and edges connecting the nodes, the nodes representing agents in the network and the edges representing actions by the agents in the network.

14. The computer program product as set forth in claim 12 , wherein if the attributes at a given node are a time series, attribute vectors at each time step can be packed into a matrix of attributes X that is N by t dimensional, where t is the number of time steps in a signal and N is a number of nodes in the network.

15. The computer program product as set forth in claim 12 , wherein the changing states of agents is based on a time-series of each agent.

16. The computer program product as set forth in claim 12 , wherein learning the time-evolving states of agents using dictionary learning is performed using the attribute matrix.

17. The computer program product as set forth in claim 12 , wherein learning the time-evolving states of agents using dictionary learning is performed in parallel using submatrices of the attribute matrix.

18. A computer implemented method for inferring network dynamics and their sources within a network, the method comprising an act of:
causing one or more processers to execute instructions encoded on a non-transitory computer-readable medium, such that upon execution, the one or more processors perform operations of:
generating a vector representation based on states of agents in a network, the vector representation including attribute vectors that correspond to the states of the agents in the network;
generating an attribute matrix based on the changing states of agents by packing the attribute vectors at each time step into the attribute matrix;
learning time-evolving states of agents using dictionary learning, wherein learning time-evolving states of agents using dictionary learning is performed through non-negative matrix factorization (NMF);
identifying influential source agents in the network by performing dimensionality reduction on the attribute matrix; and
executing an action based on the identity of the influential source agents.

19. The computer implemented method as set forth in claim 18 , wherein in generating a vector representation based on the states of agents in a network, the network includes nodes and edges connecting the nodes, the nodes representing agents in the network and the edges representing actions by the agents in the network.

20. The computer implemented method as set forth in claim 18 , wherein if the attributes at a given node are a time series, attribute vectors at each time step can be packed into a matrix of attributes X that is N by t dimensional, where t is the number of time steps in a signal and N is a number of nodes in the network.

21. The computer implemented method as set forth in claim 18 , wherein the changing states of agents is based on a time-series of each agent.

22. The computer implemented method as set forth in claim 18 , wherein learning the time-evolving states of agents using dictionary learning is performed using the attribute matrix.

23. The computer implemented method as set forth in claim 18 , wherein learning the time-evolving state

Assignees

• Hrl Lab Llc

Inventors

• Munn Steven J
• Ni Kang-Yu
• Xu Jiejun

Classifications

• H04L41/046

  comprising network management agents or mobile agents therefor · CPC title

• G06F9/5077

  Logical partitioning of resources; Management or configuration of virtualized resources (specific details on emulation or internal functioning of virtual machines G06F9/455) · CPC title

• H04L43/065

  related to network devices · CPC title

• H04L41/0893

  Assignment of logical groups to network elements · CPC title

• G06F17/16 Primary

  Matrix or vector computation {, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization (matrix transposition G06F7/78)} · CPC title