Encoding of data for processing in a spatial and temporal memory system

• US 8,645,291 B2
• Filed: 08/25/2011
• Issued: 02/04/2014
• Est. Priority Date: 08/25/2011
• Status: Active Grant

• Current Assignee: Numenta, Inc.
• Original Assignee: Numenta, Inc.
• Inventors: Hawkins, Jeffrey C., Marianetti, Ronald II, Raj, Anosh
• Primary Examiner(s): Chen, Alan
• Application Number:
  US13/218,170
• Publication Number:
  US 20130054495A1
• Time in Patent Office: 894 Days
• Field of Search: None
• US Class Current: 706/12
• CPC Class Codes: G06F 13/385 for adaptation of a particular data processing system to different peripheral devices
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 9/3867 using instruction pipelines
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 20/20 Ensemble learning

Abstract

A spatial and temporal memory system (STMS) processes input data to detect whether spatial patterns and/or temporal sequences of spatial patterns exist within the data, and to make predictions about future data. The data processed by the STMS may be retrieved from, for example, one or more database fields and is encoded into a distributed representation format using a coding scheme. The performance of the STMS in predicting future data is evaluated for the coding scheme used to process the data as performance data. The selection and prioritization of STMS experiments to perform may be based on the performance data for an experiment. The best fields, encodings, and time aggregations for generating predictions can be determined by an automated search and evaluation of multiple STMS systems.

Claims

1. A method of processing data, comprising

retrieving at least one field of data from one or more data sources including a plurality of records of data;

preprocessing one or more entries in the at least one field of data;

converting, at an encoder, the at least one field of data into encoded input data, the encoded input data comprising one or more vectors in a distributed representation format corresponding to the preprocessed one or more entries; and

detecting spatial patterns or temporal sequences of spatial patterns in the encoded input data at a processing node.

2. The method of claim 1, further comprising configuring the encoder and the processing node to perform conversion and processing according to configuration information.

3. The method of claim 2, wherein the one or more data sources comprises one or more of:

a database, an external data source, and a data stream.

4. The method of claim 3, further comprising supplanting or adding data from the external data source to a field of data of the database.

5. The method of claim 1, wherein converting the at least one field comprises:

converting an entry of a first data field into first encoded data in a distributed representation format;

converting an entry of a second data field into second encoded data in a distributed representation format; and

concatenating the first encoded data and the second encoded data to form the encoded input data.

6. The method of claim 1, wherein the at least one field of data includes an entry represented by a scalar value, and wherein the conversion comprises determining whether the scale value is within one or more buckets, wherein each bucket defines a numerical range.

7. The method of claim 6, wherein at least two buckets have overlapping numerical ranges.

8. The method of claim 1, wherein the preprocessing values comprises performing time aggregation on a subset of the values to generate the preprocessed values.

9. A system for processing data, the system comprising:

a data retrieving module configured to retrieve at least one field of data from one or more data sources including a plurality of records of data;

a preprocessor configured to preprocess one or more entries in the at least one field of data;

an encoder configured to convert the at least one field of data into encoded input, the encoded input data comprising one or more vectors in a distributed representation format corresponding to the preprocessed one or more entries; and

a processing node configured to detect spatial patterns or temporal sequences of spatial patterns in the encoded input data.

10. The system of claim 9, further comprising a configuration module configured to configure the encoder and the processing node to perform conversion and processing according to configuration information.

11. The system of claim 10, wherein the one or more data sources comprises one or more of:

a database, an external data source, and a data stream.

12. The system of claim 11, further comprising supplanting or adding data from the external data source to a field of data of the database.

13. The system of claim 9, wherein converting the at least one field comprises:

converting an entry of a first data field into first encoded data in a distributed representation format;

converting an entry of a second data field into second encoded data in a distributed representation format; and

concatenating the first encoded data and the second encoded data to form the encoded input data.

14. The system of claim 9, wherein the at least one field of data includes an entry represented by a scalar value, and wherein the conversion comprises determining whether the scale value is within one or more buckets, wherein each bucket defines a numerical range.

15. The system of claim 14, wherein at least two buckets have overlapping numeral ranges.

16. The system of claim 9, wherein the preprocessing values comprises performing time aggregation on a subset of the values to generate the preprocessed values.

17. A non-transitory computer-readable storage medium storing executable computer program instructions for processing data, the instructions comprising instructions for:

retrieving at least one field of data from one or more data sources including a plurality of records of data;

preprocessing one or more entries in the at least one field of data;

converting, at an encoder, the at least one field of data into encoded input data, the encoded input data comprising one or more vectors in a distributed representation format corresponding to the preprocessed one or more entries; and

detecting spatial patterns or temporal sequences of spatial patterns in the encoded input data at a processing node.

18. The non-transitory computer-readable storage medium of claim 17, further comprising configuring the encoder and the processing node to perform conversion and processing according to configuration information.

19. The non-transitory computer-readable storage medium of claim 17, wherein the at least one field of data includes an entry represented by a scalar value, and wherein the conversion comprises determining whether the scale value is within one or more buckets, wherein each bucket defines a numerical range.

20. The non-transitory computer-readable storage medium of claim 19, wherein at least two buckets have overlapping numerical ranges.