公开(公告)号：US20220091912A1

公开(公告)日：2022-03-24

申请号：US17026372

申请日：2020-09-21

Applicant: International Business Machines Corporation

Inventor： Larisa Shwartz ,  Qing Wang ,  Jinho Hwang ,  Srinivasan Parthasarathy ,  Michael Elton Nidd ,  Frank Bagehorn ,  Ota Sandr ,  Tomas Ondrej ,  Altynbek Orumbayev ,  Jakub Krchak ,  Michal Mylek

IPC: G06F9/54 ,  G06F9/46 ,  G06F16/28 ,  G06N20/00

Abstract: Embodiments of the present invention provide methods, computer program products, and systems. Embodiments of the present invention can dynamically determine one or more endpoints to fulfill a user request. Embodiments of the present invention can select the dynamically determined one or more endpoints as the one or more endpoints that fulfill the user request. Embodiments of the present invention can execute the selected one or more endpoints to fulfill the user request.

公开(公告)号：US20210342132A1

公开(公告)日：2021-11-04

申请号：US16863001

申请日：2020-04-30

Applicant: International Business Machines Corporation

Inventor： Srinivasan Parthasarathy ,  Fabio A. Oliveira ,  Sushma Ravichandran

IPC: G06F8/41 ,  G06F11/34 ,  G06F11/30 ,  G06N20/00 ,  G06N5/04

Abstract: A system, computer program product, and method are provided for orchestrating a multi objective optimization of an application. A set of two or more key performance indicators (KPIs) and one or more parameters associated with the application are received. A machine learning (ML) based surrogate function learning model in combination with an acquisition function is leveraged to conduct one or more adaptive trials. Each trial consists of a specific configuration of the one or more parameters. A pareto surface of the KPIs of the application is computed based on the observations of KPI values from each adaptive trial. The pareto surface is explored and an optimal operating point is selected for the application. The application is then executed at the selected operating point.

公开(公告)号：US10917316B2

公开(公告)日：2021-02-09

申请号：US16427768

申请日：2019-05-31

Applicant: International Business Machines Corporation

Inventor： Srinivasan Parthasarathy ,  Fabio A. Oliveira ,  Sushma Ravichandran

IPC: G06F15/173 ,  H04L12/24 ,  G06F16/2455

Abstract: A specification of a topology of a microservices application is obtained as a plurality of nodes corresponding to a plurality of microservices of the microservices application. The plurality of nodes includes a root node, each of the plurality of nodes other than the root node has a timeout parameter and a retries parameter. Via constrained black box optimization, optimized values are selected for the timeout parameter and the retries parameter for each of the plurality of nodes other than the root node, subject to satisfying a specified end-to-end latency for the microservices application and minimizing an error rate for the microservices application. The microservices application is configured in accordance with the optimized values. At least one external request to the root node is responded to with the microservices application configured in accordance with the optimized values.

公开(公告)号：US10031945B2

公开(公告)日：2018-07-24

申请号：US15385085

申请日：2016-12-20

Applicant: International Business Machines Corporation

Inventor： Charu Aggarwal ,  Yanjie Fu ,  Srinivasan Parthasarathy ,  Deepak Turaga

IPC: G06F17/30

Abstract: An automated outlier detection system implements an unsupervised set of processes to determine feature subspaces from a dataset; determine candidate exploratory actions, where each candidate exploratory action is a specific combination of a feature subspace and a parameterized instance of an outlier detection algorithm; and identify a set of optimal exploratory actions to recommend for execution on the dataset from among the candidate exploratory actions. Outlier scores obtained as a result of execution of the set of optimal exploratory actions are processed to obtain one or more outlier views such that each outlier view represents a consistent characterization of outliers by each optimal exploratory action corresponding to that outlier view.

公开(公告)号：US20170228650A1

公开(公告)日：2017-08-10

申请号：US15018411

申请日：2016-02-08

Applicant: International Business Machines Corporation

Inventor： Yanjie Fu ,  Srinivasan Parthasarathy ,  Deepak Turaga

IPC: G06N5/04 ,  G06N99/00

CPC classification number: G06N5/04 ,  G06N99/005

Abstract: A computer program product is provided for estimating algorithm run times given parameters of the algorithm, specifications of an architecture on which the algorithm will execute and dimensions of a data set which will be input into the algorithm. The computer program product includes instructions to cause a processing circuit to create training data sets, generate run time data of executions of instances of the algorithm on the architecture for each training data set, identify model-usable features, generate a map of the model-usable features to an expression of the run time data and iteratively tuning the model-usable features toward improving map accuracy until a target map accuracy is achieved, develop a predictive model based on iteratively tuned versions of the model-usable features and estimate a run time of an execution of the algorithm on a new data set and on a new architecture using the predictive model.

公开(公告)号：US20160154719A9

公开(公告)日：2016-06-02

申请号：US13972566

申请日：2013-08-21

Applicant: International Business Machines Corporation

Inventor： Wei Fan ,  Nagui Halim ,  Mark C. Johnson ,  Srinivasan Parthasarathy ,  Deepak S. Turaga ,  Olivier Verscheure

IPC: G06F11/27

CPC classification number: G06F11/27 ,  G01R31/2894

Abstract: The present invention describes a method and system for optimizing a test flow within each ATE (Automated Test Equipment) station. The test flow includes a plurality of test blocks. A test block includes a plurality of individual tests. A computing system schedule the test flow based one or more of: a test failure model, test block duration and a yield model. The failure model determines an order or sequence of the test blocks. There are at least two failure models: independent failure model and dependant failure model. The yield model describes whether a semiconductor chip is defective or not. Upon completing the scheduling, the ATE station conducts tests according to the scheduled test flow. The present invention can also be applied to software testing.

公开(公告)号：US20160110362A1

公开(公告)日：2016-04-21

申请号：US14518506

申请日：2014-10-20

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION

Inventor： Udayan Khurana ,  Srinivasan Parthasarathy ,  Venkata N. Pavuluri ,  Deepak S. Turaga ,  Long H. Vu

IPC: G06F17/30

CPC classification number: G06F16/242 ,  G06F16/245 ,  G06F16/24553 ,  G06F16/248 ,  G06F16/3323

Abstract: Embodiments relate to analyzing dataset. A method of analyzing data is provided. The method obtains a description of a dataset. The method automatically generates a plurality of analysis options from the description of the dataset. The method generates a plurality of queries based on the analysis options. The method deploys the queries on the dataset to build a plurality of statistical models from the dataset.

Abstract translation: 实施例涉及分析数据集。 提供了一种分析数据的方法。 该方法获取数据集的描述。 该方法根据数据集的描述自动生成多个分析选项。 该方法基于分析选项生成多个查询。 该方法在数据集上部署查询以从数据集构建多个统计模型。

公开(公告)号：US20220156065A1

公开(公告)日：2022-05-19

申请号：US16951500

申请日：2020-11-18

Applicant: International Business Machines Corporation

Inventor： Michael Elton Nidd ,  Altynbek Orumbayev ,  Jinho HWANG ,  Larisa Shwartz ,  Jakub Krchak ,  Qing Wang ,  Frank Bagehorn ,  Ota Sandr ,  Tomas Ondrej ,  Srinivasan Parthasarathy ,  Michal Mylek

IPC: G06F8/77 ,  G06F8/71 ,  G06Q10/10 ,  G06Q30/00 ,  G06N20/00

Abstract: In an approach to linking operational data with issues, a new event is received. The new event is associated to a story, where the story is related to an identified problem within the system, and further where the new event is associated with the story using machine learning techniques. The story is associated to related change requests based on a similarity between the story and related change requests, where the similarity between the story and the related change requests is associated using the machine learning techniques. A cost is calculated for the story. Responsive to associating the new event with a specific change request, the priority of the specific change request is updated based on the cost for the story.

公开(公告)号：US20220035692A1

公开(公告)日：2022-02-03

申请号：US16945954

申请日：2020-08-03

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION

Inventor： Qing Wang ,  Larisa Shwartz ,  Jinho Hwang ,  Srinivasan Parthasarathy ,  Michael Elton Nidd ,  Frank Bagehorn ,  Jakub Krchák ,  Tomás Ondrej ,  Altynbek Orumbayev ,  Michal Mýlek ,  Ota Sandr

IPC: G06F11/07 ,  G06F17/16

Abstract: An approach to recommending corrective action to computing system event errors. The approach may include generating a textual description of an event error. The approach may include transforming the textual description into feature vectors with a domain-specific word embedding module. The approach may also include generating a recommendation to correct the event error based on an analysis of the feature vectors. Additionally, the recommendation may be presented for verification.

公开(公告)号：US20200265071A1

公开(公告)日：2020-08-20

申请号：US16280528

申请日：2019-02-20

Applicant: International Business Machines Corporation

Inventor： Srinivasan Parthasarathy ,  Tejaswini Pedapati

IPC: G06F16/28 ,  G06F16/25 ,  G06N20/00 ,  G06F9/50

Abstract: Computerized interactive feature visualization is carried out on a data set—a plurality of insight classes rank a plurality of features of the data set. Via a computerized user interface, user feedback is obtained based on the interactive feature visualization—a user selects and ranks a subset of the features. At least one transformation function is applied to at least one feature of the subset of features selected by the user, to automatically construct, with a computer, at least one additional feature for the data set. The data set with the at least one additional feature is a transformed data set. In some cases, a supervised task is carried out on the final data set; accuracy of a machine learning system implementing the at least one supervised task can be enhanced by the at least one additional feature, and/or a physical system can be controlled based on results of the at least one supervised task.