Pivotal Software, Inc.

Verint Systems, Inc.

Primesense Ltd

Apple Inc.

Michael Brand

Eran Guendelman

Amiad Gurman

Tomer Yanir

David Russell Brown

Kee Siong Ng

Lyndon John Adams

Chunsheng Fang

Florian Schoppmann

Jarrod James Vawdrey

Emily Kawaler

Ofer Shochet

Ido Shapira

Galia Zacay

Yaron Gvili

Alexander Sokolovsky

Sagi Nachum

Dror Daniel Ziv

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for routing events of an event stream in a stream processing system. One of the methods includes receiving, by a router, an event stream of events; identifying, for each event, by the router, a respective partition of context data that includes context data related to the event and providing the event to a respective local modeler that stores the partition of context data identified for the event in operational memory of the local modeler; processing, by each local modeler, events received from the router and aggregating information associated with each event to generate aggregated information; providing, by one or more of the local modelers, to a central modeler, the respective aggregated information; and determining, by the central modeler, a plurality of parameters of a machine learning model using the received aggregated information.

Compute intensive stream processing with context data routing

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for stream processing. One method includes receiving an event stream of events by a first plurality of local modelers of a stream processing system. Each local modeler processes a portion of received events of the event stream according to a first set of operations, the operations including aggregating information associated with each event to generate aggregated information. One or more local modelers provide, to a first central modeler executing on the system, the respective aggregated information generated by one or more of the local modelers. A set of parameters of a respective machine learning model is determined using the received aggregated information.

Compute intensive stream processing

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for stream processing. One method includes receiving an event stream of first events by a first plurality of first local modelers of a stream processing system. Each local modeler processes a portion of received events of the event stream according to a first set of operations, the operations including aggregating information associated with each event to generate first aggregated information. A second plurality of second local modelers similarly generates second aggregated information from an event stream of second events. First and second local modelers provide, to a first central modeler, first and second aggregated information. A set of parameters of a respective machine learning model is determined by the first central modeler using the received aggregated information.

Stream processing with multiple connections between local and central modelers

Methods, systems, and apparatus, including computer programs encoded on computer storage media, for routing events of an event stream. One of the methods includes operations of receiving, by a router, events of an event stream; providing each event, by the router, to a respective local modeler selected by the router according to an initial routing strategy, the respective local modeler being selected from multiple local modelers; aggregating, by each local modeler in parallel, information associated with each event received by the local modeler to generate aggregated information; providing, to a central modeler, the aggregated information generated by the one or more local modelers; determining, by the central modeler, parameters of a machine learning model using the aggregated information received by the central modeler and generating an updated routing strategy based on the parameters of the machine learning model; and providing the updated routing strategy to the router.

Stream processing with dynamic event routing

A method determines the connectivity components defined by a set of relations over a set of data elements. For each first data element of a selected subset of data elements, a second data element that is linked to the first data element by a path of relations is selected as its representative, using a randomization process. A new set of relations is created by replacing each first data element of the subset by its representative in at least part of the set of relations.

In-database connectivity components analysis of data

A method of operating a speech processing system is provided. The method includes translating a portion of a speech record into a plurality of possible words associated with a plurality of contexts, and determining a plurality of correctness values based on a plurality of probabilities that each of the plurality of possible words is correct for each of the plurality of contexts. The method also includes determining which of the plurality of possible words is a correct translation of the portion of the speech record based on the plurality of correctness values.

System and method for context sensitive inference in a speech processing system

A method of database indexing is disclosed. Two functions (f and g) from row values to row number values are set. The functions are utilized to determine a row number in a database column containing a target search value, wherein the target search value comprises a search value being sought in the database column. A candidate row number variable is set initially to the function g of the target search value by a processor. Iteratively the following is performed: a current value of the candidate row number variable is used as an address to read a value in a corresponding row in the database column, and the current value of the candidate row number variable is updated to the function f of the most recently read value in a corresponding row in the database column.

Queries for thin database indexing

A method of indexing a database column is disclosed. A permutation function f is determined. A shortcut that connects two non-adjacent elements of a permutation cycle based on f is created. A traversal of the permutation cycle without the shortcut comprises starting at a first element of the permutation cycle by updating a candidate row number to an initial value, wherein the initial value comprises a value stored in the database column. Traversing through one element of the permutation cycle comprises reading a row value in the database column using the candidate row number and then updating the candidate row number with function f of the most recently read row value. The step of traversing through one element of the permutation cycle is repeated until the most recently read row value is equal to the initial value. Creating the shortcut comprises storing a relation between the two non-adjacent elements.

Thin database indexing

A method for processing data includes receiving a depth map of a scene containing a humanoid form. The depth map is processed so as to identify three-dimensional (3D) connected components in the scene, each connected component including a set of the pixels that are mutually adjacent and have mutually-adjacent depth values. Separate, first and second connected components are identified as both belonging to the humanoid form, and a representation of the humanoid form is generated including both of the first and second connected components.

Analysis of three-dimensional scenes

A computer-implemented method for processing a plurality of data items includes defining a set of one or more categories having a corresponding set of conditions that associate the data items with the categories. A sub-categorization request, requesting to divide a category from among the categories into lower-level categories, is accepted from a user. The data items associated with the category are processed responsively to the sub-categorization request, so as to automatically suggest the lower-level categories. The automatically-suggested lower-level categories are presented to the user, and direction with respect to the automatically-suggested lower-level categories is accepted from the user. A hierarchical structure representing the categories is constructed responsively to the direction, by dividing the category into the lower-level categories. Output based on the hierarchical structure is presented to the user.

Root cause analysis using interactive data categorization

A method for automatically providing a plurality of additional database query terms comprising receiving a first query term from a user, receiving a plurality of characters from the user, wherein the plurality of characters is only a portion of a second query term, and selecting a set of records from a database based on the query term, wherein the database comprises records which comprise text translated from audio. The method also determines a plurality of additional query terms based on the plurality of characters, and, for at least one of the plurality of additional query terms, processes at least a portion of the set of records to determine a relevance of the additional query term. Finally, the method includes displaying at least one of the plurality of additional query terms to the user for selection based on the relevance of at least one of the plurality of additional query terms.

Growing community of inventors

Bentleigh East, Australia

Michael Brand