EXTENSIBLE HIERARCHICAL TEMPORAL MEMORY BASED SYSTEM

• US 20070276774A1
• Filed: 01/11/2007
• Published: 11/29/2007
• Est. Priority Date: 02/10/2006
• Status: Abandoned Application

• Current Assignee: Numenta, Inc.
• Original Assignee: Numenta, Inc.
• Inventors: Hawkins, Jeffrey, George, Dileep, Astier, Frank, Ahmad, Subutai
• Application Number:
  US11/622,447
• Publication Number:
  US 20070276774A1
• Time in Patent Office: Days
• US Class Current: 706/12
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 3/04 Architecture, e.g. interconnection topology
• CPC Class Codes: G06N 3/045 Combinations of networks
• CPC Class Codes: G06N 3/047 Probabilistic or stochastic networks
• CPC Class Codes: G06N 3/049 Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
• CPC Class Codes: G06N 3/08 Learning methods
• CPC Class Codes: G06N 5/02 Knowledge representation; Symbolic representation
• CPC Class Codes: G06N 7/01 Probabilistic graphical models, e.g. probabilistic networks

Abstract

A hierarchy of computing modules is configured to learn a cause of input data sensed over space and time. The hierarchy is further configured to determine a cause of novel sensed input data dependent on the learned cause. Further, the hierarchy has at least one interface associated with an operation of the hierarchy, where the operation is extensible via the interface.

Claims

1. A system, comprising:

a hierarchy of computing modules configured to learn a cause of input data sensed over space and time, the hierarchy further configured to determine a cause of novel sensed input data dependent on the learned cause, the hierarchy having at least one interface associated with an operation of the hierarchy,

wherein the operation is extensible via the interface.

2. The system of claim 1, wherein the hierarchy comprises a first level of a plurality of computing modules and a second level of at least one computing module, and wherein an input range of the computing module in the second level is greater than an input range of a computing module in the first level.

3. The system of claim 1, wherein the hierarchy is further configured to determine the cause of the novel sensed input data dependent on a computed inference of the cause passed from the first level to the second level.

4. The system of claim 3, wherein the hierarchy is further configured to determine the cause of the novel sensed input data dependent on a prediction vector of the cause passed from the second level to the first level.

5. The system of claim 1, wherein, after the operation has been extended, the extended operation is invocable dependent on a priority level associated with the interface.

6. The system of claim 1, wherein, after the operation has been extended, the extended operation is invocable dependent on a time designation associated with the interface.

7. The system of claim 1, wherein the interface comprises at least one of a sensor interface, an effector interface, a link interface, a supervisor process interface, a router interface, and a learning algorithm interface.

8. A computer-implemented method, comprising:

running a network of a hierarchy of computing modules configured to learn a cause of a first set of input data received over space and time and further configured to determine a cause of a second set of input data dependent on the learned cause, wherein running the network comprises;

processing a plurality of entities of the network according to a priority associated with each of the entities, wherein at least one of the entities is extensible.

9. The computer-implemented method of claim 8, wherein the at least one entities is implemented using a base class.

10. The computer-implemented method of claim 9, wherein running the network comprises:

compiling source code for the network, wherein the source code comprises source code for a subclass extending the base class.

11. The computer-implemented method of claim 9, wherein running the network comprises:

dynamically linking a library during operation of the network, wherein the library comprises code for extending the base class.

12. The computer-implemented method of claim 11, wherein the library comprises a plug-in library.

13. The computer-implemented method of claim 8, wherein the at least one entity is associated with handling input data sensed by the network.

14. The computer-implemented method of claim 8, wherein the at least one entity is associated with handling an output of the network.

15. The computer-implemented method of claim 8, wherein the at least one entity is associated with handling message passing between a first level of the hierarchy and a second level of the hierarchy.

16. The computer-implemented method of claim 8, wherein the at least one entity is associated with a learning algorithm used by the network to learn the cause of the first set of input data.

17. The computer-implemented method of claim 8, wherein the priority of an entity is dependent on a time stamp associated with the entity.

18. A system, comprising:

an HTM network having a plurality of interfaces, at least one of which is implemented in software using a base class, wherein the at least one interface is extensible using a subclass extending the base class, and

wherein the HTM network, using the extended interface, is configured to at least one of learn a cause of first set of input data received over space and time and determine a cause of a second set of input data dependent on the learned cause.

19. A computer-readable medium having instructions stored therein that are executable by a processor, the instructions comprising instructions to:

run a network of a hierarchy of computing modules configured to learn a cause of a first set of input data received over space and time and further configured to determine a cause of a second set of input data dependent on the learned cause, wherein the instructions to run the network comprise instructions to;

process a plurality of entities of the network according to a priority associated with each of the entities, wherein at least one of the entities is extensible.

20. The computer-readable medium of claim 19, wherein the at least one entities is implemented using a base class.

21. The computer-readable medium of claim 20, wherein the instructions to run the network comprise instructions to:

compile source code for the network, wherein the source code comprises source code for a subclass extending the base class.