Numenta, Inc.

Charles Curry

Jeffrey Charles Hawkins

Subutai Ahmad

Dileep George

Frank Emmanuel Astier

Robert Gilchrist Jaros

Anosh Raj

Giyora Sayfan

A temporal pooler for a Hierarchical Temporal Memory network is provided. The temporal pooler is capable of storing information about sequences of co-occurrences in a higher-order Markov chain by splitting a co-occurrence into a plurality of sub-occurrences. Each split sub-occurrence may be part of a distinct sequence of co-occurrences. The temporal pooler receives the probability of spatial co-occurrences in training patterns and tallies counts or frequency of transitions from one sub-occurrence to another sub-occurrence in a connectivity matrix. The connectivity matrix is then processed to generate temporal statistics data. The temporal statistics data is provided to an inference engine to perform inference or prediction on input patterns. By storing information related to a higher-order Markov model, the temporal statistics data more accurately reflects long temporal sequences of co-occurrences in the training patterns.

Hierarchical temporal memory system with higher-order temporal pooling capability

A system for implementing a hierarchical temporal memory (HTM) network using a plugin infrastructure. The plugin infrastructure registers the plugins to be used in instantiating the HTM network. The plugin may include one or more functions for creating one or more components of the HTM network in a runtime engine. The plugin is associated with a component specification describing the components of the HTM network created by invoking the functions of the plugin. After the plugin is registered, the plugin infrastructure allows functions of the plugin to be invoked to instantiate The HTM network on a runtime engine. After the HTM network is instantiated, the runtime engine may run the instance of the HTM network to learn and infer the causes of input data. The system may also include one or more external programs to provide various supporting operations associated with the runtime engine by referencing the component specification. The supporting operations that can be performed by the external programs include, among others, validating a netlist defining the structure of the HTM network, build help information for the component, and generating a graphical user interface associated with the HTM network.

Plugin infrastructure for hierarchical temporal memory (HTM) system

Growing community of inventors

Fremont, CA, United States of America

Charles Curry