Sas Institute Inc.

Steel King Industries, Inc.

Cisco Technology, Inc.

Steven E Krueger

Joshua David Griffin

Brian Payton Bowman

Lawrence Edmund Lewis

Xindian Long

Yan Xu

Scott Russell Pope

Xingqi Du

Gordon Lyle Keener

Liping Cai

Jorge Manuel Gomes Da Silva

Ralph Walter Abbey

Mohammadreza Nazari

Amirhassan Fallah Dizche

Richard Todd Knight

Chih-Wei Ho

Gary J Rosenberg

Stacey Michelle Christian

Steven Daniel Miles

Jayce S Anderson

William R Sniff

A system, method, and computer-program product includes selecting, by a controller node, a plurality of candidate hyperparameter search points from a hyperparameter search space; instructing, by the controller node, one or more worker nodes to concurrently train a plurality of machine learning models for a target number of epochs using the plurality of candidate hyperparameter search points; identifying, by the controller node, a collection of intermediate candidate hyperparameter search points that outperform a remainder of the plurality of candidate hyperparameter search points by evaluating a performance of the plurality of machine learning models after training for the target number of epochs; and performing, by the controller node, a crossover operation with the collection of intermediate candidate hyperparameter search points to identify one or more new candidate hyperparameter search points to test in the hyperparameter search space.

Systems and methods for parallel exploration of a hyperparameter search space

A system, method, and computer-program product includes commencing a parent computer process based on receiving a request to perform an analytical operation on one or more datasets, commencing at least one child computer process that is launched by the parent computer process when the parent computer process initiates an execution of the analytical operation on the one or more datasets, transmitting, by the at least one child computer process, a request to the parent computer process to retrieve the one or more datasets, writing, by the parent computer process, the one or more datasets to a cross-process queue based on the parent computer process receiving the requests, reading, by the at least one child computer process, the one or more datasets from the cross-process queue, and executing, using an analytical application executing on the least one child computer process, the analytical operation based on the one or more datasets.

Systems and methods for executing an analytical operation across a plurality of computer processes

A system, method, and computer-program product includes computing, by a controller node, a hyperparameter search space for a plurality of hyperparameters; selecting, by the controller node, a plurality of hyperparameter search points from the hyperparameter search space; instructing, by the controller node, one or more worker nodes to concurrently train a plurality of machine learning models based on the plurality of hyperparameter search points; receiving, from the one or more worker nodes, a plurality of performance metrics that measure a performance of the plurality of machine learning models; determining, by the controller node, one or more sets of optimal hyperparameter values based on the plurality of performance metrics; and outputting, by the controller node, the one or more sets of optimal hyperparameter values.

A system, method, and computer-program product includes receiving, by a worker process, a plurality of chunks of data from a client process; deriving, by the worker process, an input pattern for feeding the plurality of chunks of data to a machine learning model; caching, by the worker process, a subset of data elements of the plurality of chunks of data specified by the input pattern based on a data caching policy; and training the machine learning model by feeding the subset of data elements cached by the worker process and a remainder of data elements in the plurality of chunks of data when requested by the input pattern.

Systems and methods for multi-language training of machine learning models

A system, method, and computer-program product includes receiving, by a controller node, a request to execute a client process associated with a first programming language and a plurality of threads; launching, by the controller node, a plurality of multi-language worker processes based on a number of threads associated with the client process; and instructing, by the controller node, the plurality of multi-language worker processes to execute the plurality of threads associated with the client process.

A system, method, and computer-program product includes selecting, by a controller node, a plurality of hyperparameter search points from a hyperparameter search space; instructing, by the controller node, one or more worker nodes to concurrently train a plurality of machine learning models for a target number of epochs using the plurality of hyperparameter search points; receiving, from the one or more worker nodes, a plurality of performance metrics that measure a performance of the plurality of machine learning models during the target number of epochs; and removing, by the controller node, one or more underperforming hyperparameter search points from the plurality of hyperparameter search points according to a pre-defined performance metric ranking criterion associated with the plurality of performance metrics.

An apparatus including a processor to index data records within a data cell, wherein for each data record, the processor retrieves data values from first and second data fields; determines whether the first and second data fields store unique data values; in response to the first data field storing a unique data value, adds an identifier of the data record to a first unique values index, in response to the second data field storing a unique data value, adds the identifier to a second unique values index, wherein identifiers of data records within the unique values indexes are ordered based on corresponding unique data values; and generates an indication of ranges of data values of the first and second data fields to enable a determination of whether a data value specified in search criteria is present within at least the data cell.

Distributed data set indexing

An apparatus includes a processor component caused to: retrieve metadata of organization of data within a data set, and map data of organization of data blocks within a data file; receive indications of which node devices are available to perform a processing task with a data set portion; and in response to the data set including partitioned data, compare the quantities of available node devices and of the node devices last involved in storing the data set. In response to a match, for each map data map entry: retrieve a hashed identifier for a data sub-block, and a size for each of the data sub-blocks within the corresponding data block; divide the hashed identifier by the quantity of available node devices; compare the modulo value to a designation assigned to each of the available node devices; and provide a pointer to the available node device assigned the matching designation.

Distributed data set storage and retrieval

An apparatus including a processor to: index multiple data records within a data cell by first and second data fields in a single read pass through the data cell; wherein for each data record within the first data cell, the processor is to retrieve data values from the first and second data fields, search a first binary tree to determine whether the data value from the first data field comprises a unique value, and add the data value to the first binary tree if it is unique, and search a second binary tree to determine whether the data value from the second data field comprises a unique value, and add the data value to the second binary tree if it is unique; and generate a first and second unique values indexes of identifiers of the data records associated with the unique data values within the first and second binary trees.

An apparatus including a processor to receive search criteria including a data value for a search within a data field; in response to the receipt of the query instructions, and for each data cell within a super cell, perform the specified search by comparing the data value to ranges of values indicated in a corresponding cell index to determine whether the data cell includes a data record meeting the search criteria, and in response to a determination that the data cell includes such a data record, use a unique values index in the cell index to search the data records of the data cell to identify one or more data records meeting the search criteria; and in response to identifying at least one data record meeting the search criteria, provide an indication that at least the data cell includes at least one data record meeting the search criteria.

An apparatus including a processor to: receive search criteria including a data value; in response to receiving the search criteria, generate a hash value from the data value of the search criteria, and for each data cell of a super cell, compare the hash value to hash values within a hash values vector in the corresponding cell index to determine whether the data cell includes at least one data record meeting the search criteria, and in response to determining that the data cell includes at least one of such data record, search the data records to identify one or more data records meeting the search criteria; and in response to identifying at least one data record within at least one data cell of the super cell meeting the search criteria, provide results data indicative of the super cell including at least one of such data record.

Growing community of inventors

Raleigh, NC, United States of America

Steven E Krueger