DIRECTED BEHAVIOR USING A HIERARCHICAL TEMPORAL MEMORY BASED SYSTEM

• US 20070192268A1
• Filed: 01/11/2007
• Published: 08/16/2007
• Est. Priority Date: 02/10/2006
• Status: Abandoned Application

• Current Assignee: Numenta, Inc.
• Original Assignee: Numenta, Inc.
• Inventors: Hawkins, Jeffrey, George, Dileep
• Application Number:
  US11/622,448
• Publication Number:
  US 20070192268A1
• Time in Patent Office: Days
• US Class Current: 706/52
• 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, and is further configured to determine a cause of novel sensed input data dependent on the learned cause. In some cases, the cause of the input data is visual or acoustic behavior (caused, for example, by a motor or processing circuitry). As such, the hierarchy is further configured to associate the behavior with a mechanism that generates the behavior, thereby enabling the hierarchy to later predict and cause the behavior to occur.

Claims

1. A method of performing operations, comprising:

receiving data caused by a manifestation of a behavior;

learning a cause of the received data and forming a representation of the behavior;

associating the representation with a mechanism that generates the behavior; and

causing the behavior to occur using the mechanism.

2. The method of claim 1, wherein the manifestation of the behavior comprises at least one of visual behavior and acoustic behavior.

3. The method of claim 1, wherein the mechanism comprises at least one of a motor and processing circuitry.

4. The method of claim 1, further comprising:

receiving new data;

predicting an occurrence of the behavior dependent on the new data, wherein causing the behavior is dependent on the prediction.

5. 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,

wherein the cause of the input data is motor behavior, and

wherein the hierarchy is further configured to associate the motor behavior with a mechanism that generates the motor behavior.

6. The system of claim 5, wherein the hierarchy is further configured to cause the motor behavior to occur using the mechanism.

7. The system of claim 5, wherein the hierarchy is further configured to predict an occurrence of the motor behavior dependent on the novel sensed input data.

8. The system of claim 7, wherein the hierarchy is further configured to cause the motor behavior to occur dependent on the prediction.

9. The system of claim 5, wherein at least one of the computing module in the first level and the computing module in the second level is implemented at least in part in at least one of software and hardware.

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

receive data caused by a manifestation of a behavior;

learn a cause of the received data and forming a representation of the behavior;

associate the representation with a mechanism that generates the behavior; and

cause the behavior to occur using the mechanism.

11. The computer-readable medium of claim 10, wherein the manifestation of the behavior comprises at least one of visual behavior and acoustic behavior.

12. The computer-readable medium of claim 10, wherein the mechanism comprises a at least one of a motor and processing circuitry.

13. The computer-readable medium of claim 10, further comprising instructions to:

receive new data;

predict an occurrence of the behavior dependent on the new data, wherein at least one of the instructions to cause the behavior is dependent on the prediction.

14. A computer-implemented method of performing operations, comprising:

learning a cause of a first set of input data, the cause being at least one of visual and acoustic behavior;

forming a representation of the behavior;

associating the representation with a mechanism used to cause the behavior;

receiving a second set of input data;

predicting an occurrence of the behavior dependent on the second set of input data; and

causing the behavior to occur dependent on the prediction.

15. The computer-implemented method of claim 14, wherein the mechanism comprises at least one of a motor and processing circuitry.

16. An apparatus, comprising:

a processor;

a memory operatively connected to the processor; and

instructions stored in the memory and executable by the processor, the instructions to;

represent at least part of an HTM,

learn a cause of sensed input data received by the HTM, the cause being a behavior generated by the apparatus,

form a representation of the behavior in the HTM, and

associate the representation with a mechanism in the apparatus capable of generating the behavior.

17. The apparatus of claim 16, the instructions further to predict in and cause by the HTM an occurrence of the behavior.

18. The apparatus of claim 16, the instructions further to cause the behavior in combination with causing another behavior.

19. The apparatus of claim 18, wherein the combination of behaviors has not previously been caused by the apparatus independent of the HTM.

20. The apparatus of claim 16, wherein the sensed input data received by the HTM comprises spatial patterns received over time.