Sas Institute Inc.

Scott Russell Pope

Joshua David Griffin

Yan Xu

Georges H Guirguis

Steven E Krueger

Lawrence Edmund Lewis

Xindian Long

Xingqi Du

Liping Cai

Oliver Schabenberger

Xinmin Wu

Patrick Nathan Koch

Tao Wang

Brett Alan Wujek

Xiangqian Hu

Steven Joseph Gardner

Oleg Borisovich Golovidov

Guixian Lin

Guy Blanc

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.

Systems and methods for parallel exploration of a hyperparameter search space

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.

A computing device computes a quantile value for a variable value extracted from an event block object by computing a bin number for the variable value. If the computed bin number is between a before bin number and an after bin number computed for a quantile, the quantile is identified. Frequency data is updated to include the extracted variable value as a key value. A frequency value associated with the key value indicates a number of occurrences of the variable value in previously processed data. A cumulative rank value of the identified quantile is updated. A quantile adjustment value is computed based on a comparison between the variable value and a current quantile value of the identified quantile. An updated quantile value associated with the identified quantile is computed using the updated frequency data, the computed quantile adjustment value, and the updated cumulative rank value of the identified quantile.

Analytic system for streaming quantile computation

A computing device automatically selects hyperparameter values based on objective criteria to train a predictive model. Each session of a plurality of sessions executes training and scoring of a model type using an input dataset in parallel with other sessions of the plurality of sessions. Unique hyperparameter configurations are determined using a search method and assigned to each session. For each session of the plurality of sessions, training of a model of the model type is requested using a training dataset and the assigned hyperparameter configuration, scoring of the trained model using a validation dataset and the assigned hyperparameter configuration is requested to compute an objective function value, and the received objective function value and the assigned hyperparameter configuration are stored. A best hyperparameter configuration is identified based on an extreme value of the stored objective function values.

Distributed hyperparameter tuning system for machine learning

In accordance with the teachings described herein, systems and methods are provided for estimating or determining quantiles for data stored in a distributed system. In one embodiment, an instruction is received to estimate or determine a specified quantile for a variate in a set of data stored at a plurality of nodes in the distributed system. A plurality of data bins for the variate are defined that are each associated with a different range of data values in the set of data. Lower and upper quantile bounds for each of the plurality of data bins are determined based on the total number of data values that fall within each of the plurality of data bins. The specified quantile is estimated or determined based on an identified one of the plurality of data bins that includes the specified quantile based on the lower and upper quantile bounds.

Systems and methods for quantile determination in a distributed data system using sampling

In accordance with the teachings described herein, systems and methods are provided for estimating quantiles for data stored in a distributed system. In one embodiment, an instruction is received to estimate a specified quantile for a variate in a set of data stored at a plurality of nodes in the distributed system. A plurality of data bins for the variate are defined that are each associated with a different range of data values in the set of data. Lower and upper quantile bounds for each of the plurality of data bins are determined based on the total number of data values that fall within each of the plurality of data bins. The specified quantile is estimated based on an identified one of the plurality of data bins that includes the specified quantile based on the lower and upper quantile bounds.

Systems and methods for quantile determination in a distributed data system

Systems and methods for quantile estimation in a distributed data system

Systems and methods are provided for identifying data variable roles during initial data exploration. A variable type, unique data value count values, and an overflow count value are determined for a variable. The unique data value count values include a number of occurrences of each of a plurality of unique data values for the variable in a data set. The overflow count value is a number of occurrences of data values other than the plurality of unique data values for the variable in the data set. When a number of the plurality of unique data values is greater than a value for a high cardinality threshold, the variable is determined to be a high cardinality variable. When the variable is not determined to be the high cardinality variable, a class variable role is assigned to the variable. When the variable is determined to be the high cardinality variable, Whether or not the variable is a numeric variable type is determined based on the determined variable type. When the variable is determined to not be the numeric variable type, the overflow count value is compared to the unique data value count values to determine whether or not rare visible values occurred for the variable. When the determination is that rare visible values occurred for the variable, a record identifier variable role is assigned to the variable.

System and method for fast identification of variable roles during initial data exploration

Systems and methods are provided for identifying data variable rules during initial data exploration. In one example, a computer-implemented method of determining a role for a data variable is disclosed. The method comprises identifying to a plurality of data nodes a set of data records containing data values assigned to each data node, a maximum number of levels to record in a sorted data structure at the data nodes, and the data node responsible for each of a plurality of variables. The method further comprises receiving for each variable from the data node responsible for the variable a plurality of unique data values for the variable, a count for each of the unique data values and an overflow count for the variable, wherein the number of unique data values does not exceed the maximum number of levels. A role for a variable can be determined based upon the unique data values, counts and overflow count for the variable.

Growing community of inventors

Raleigh, NC, United States of America

Scott Russell Pope