International Business Machines Corporation

Peter Daniel Kirchner

Jerry M Woodall

George D Pettit

Alan C Warren

Theodore Gerard Van Kessel

Steven Lorenz Wright

Thomas N Jackson

Horst Cornelius Samulowitz

John Lawrence Freeouf

Yves Corfield Martin

Dakuo Wang

Saket Sathe

Gregory Bramble

Michael A Gaynes

Jing James Xu

Brent A Wacaser

Bei Chen

Robert L Sandstrom

Naim Moumen

Si Er Han

Ji Hui Yang

Long H Vu

Dhavalkumar C Patel

Syed Yousaf Shah

Dhirendra Patel

Xuan-Hong Dang

Joel Leonard Wolf

Chuang Gan

Jai P Menon

Jun Song Wang

Petros Zerfos

James Sargent Lipscomb

David D Nolte

Paul Borrel

Rodney T Hodgson

Jun Wang

Wesley M Gifford

Kanthi K Sarpatwar

Ronald Franklin Marks

Jaroslaw R Rossignac

Alan B Fowler

Arunima Chaudhary

Richard F Rutz

Hussam Khonkar

Michael R Melloch

Robert Saccone

Matthew F Chisholm

Eric R Fossum

Jeffrey Y Tang

Robert Howard Wolfe

James J Rosenberg

Long Vu

Dennis F Manzer

Krzysztof P Onak

Gregory Filla

To automate time series forecasting machine learning pipeline generation, a data allocation size of time series data may be determined based on one or more characteristics of a time series data set. The time series data may be allocated for use by candidate machine learning pipelines based on the data allocation size. Features for the time series data may be determined and cached by the candidate machine learning pipelines. Predictions of each of the candidate machine learning pipelines using at least the one or more features may be evaluated. A ranked list of machine learning pipelines may be automatically generated from the candidate machine learning pipelines for time series forecasting based upon evaluating predictions of each of the one or more candidate machine learning pipelines.

Automated time series forecasting pipeline generation

Computer hardware and/or software that performs the following operations: (i) assessing a performance of a plurality of unsupervised machine learning pipelines against a plurality of data sets; (ii) associating the performance with meta-features corresponding to respective pipeline/data set combinations; (iii) training a supervised meta-learning model using the associated performance and meta-features as training data; and (iv) utilizing the trained model to identify one or more pipelines for processing an input data set.

Automated unsupervised machine learning utilizing meta-learning

An embodiment of the invention may include a method, computer program product, and system for creating a data analysis tool. The method may include a computing device that generates an AI pipeline based on an input dataset, wherein the AI pipeline is generated using an Automated Machine Learning program. The method may include converting the AI pipeline to a non-native format of the Automated Machine Learning program. This may enable the AI pipeline to be used outside of the Automated Machine Learning program, thereby increasing the usefulness of the created program by not tying it to the Automated Machine Learning program. Additionally, this may increase the efficiency of running the AI pipeline by eliminating unnecessary computations performed by the Automated Machine Learning program.

Code generation for Auto-AI

Techniques regarding one or more automated machine learning processes that analyze time series data are provided. For example, one or more embodiments described herein can comprise a system, which can comprise a memory that can store computer executable components. The system can also comprise a processor, operably coupled to the memory, and that can execute the computer executable components stored in the memory. The computer executable components can comprise a time series analysis component that selects a machine learning pipeline for meta transfer learning on time series data by sequentially allocating subsets of training data from the time series data amongst a plurality of machine learning pipeline candidates.

Automated machine learning pipeline generation

A method of scheduling assignment of resources to a plurality of applications includes: determining shares of the resources assigned to each application during a first period; determining shares of the resources assigned to each application during a second period that occurs after the first period; determining an imbalance value for each application that is based on a sum of the shares assigned to the corresponding application over both periods; and considering requests of the applications for resources in an order that depends on a result of comparing the imbalance values of the applications.

Multi-platform scheduler for permanent and transient applications

Techniques for optimizing power production from photo-voltaic systems using, e.g., inductive coupling, are provided. In one aspect, a method of optimizing photo-voltaic generated power from a string of photo-voltaic devices is provided. The method includes the step of: providing corrective power to at least one photovoltaic device in the string of photo-voltaic devices to boost performance of the at least one photovoltaic device and thereby increase overall the photo-voltaic generated power from the string of photo-voltaic devices, wherein the corrective power is from about 1% to about 5%, and ranges therebetween, of the photo-voltaic generated power from the string of photo-voltaic devices. A system for optimizing photo-voltaic generated power from a string of photo-voltaic devices and a method for use thereof are also provided.

Techniques for optimizing photo-voltaic power via inductive coupling

A device for dissipating heat from a photovoltaic cell is disclosed. A first thermally conductive layer receives heat from the photovoltaic cell and reduces a density of the received heat. A second thermally conductive layer conducts heat from the first thermally conductive layer to a surrounding environment. An electrically isolating layer thermally couples the first thermally conductive layer and the second thermally conductive layer.

Thermal receiver for high power solar concentrators and method of assembly

A method of tracking the sun is disclosed. A measurement of a gravity vector is obtained in a frame of reference of a solar collector rotatable with respect to an earth center of reference. A measurement of a magnetic direction vector is obtained in the frame of reference of the solar collector. The orientation of the solar collector is determined from the obtained measurement of the gravity vector and the obtained measurement of the magnetic direction. The orientation of the solar collector is altered in order to track the sun.

Controlling a solar tracking system

A graphics processing and display system includes a coordinate sensing system for determining position and orientation of a view point reference with respect to a schematic representation of at least one three-dimensional object; a view point controller for controlling a display view point according to the determined position and orientation of the view point reference with respect to the schematic representation; and a display processor for generating pixel data representing a two-dimensional rendering of the at least one three-dimensional object according to the graphics data and display view point, wherein the two-dimensional rendering is different from the schematic representation, and wherein the pixel data generated by the display processor is output for display on a display device. Advantageously, the system of the present invention is efficient and easier to use than prior art graphics processing and display systems because the user is provided the ability to control the display view point according to the position and orientation of the view point reference with respect to the schematic representation of the displayed object.

Schematic guided control of the view point of a graphics processing and

A light responsive device (10) has a body (12) that includes a matrix comprised of Group III-V material, the matrix having inclusions (14) comprised of a Group V material contained therein. The body is responsive to a presence of a light beam that has a spatially varying intensity for modifying in a corresponding spatially varying manner a distribution of trapped photoexcited charge carriers within the body. The distribution of trapped charge carriers induces a corresponding spatial variation in at least one optical property of the Group III-V material, such as the index of refraction of the Group III-V material and/or an absorption coefficient of the Group III-V material. The Group III-V material is comprised of LTG GaAs:As or LTG AlGaAs:As. In an optical storage medium embodiment of the invention the spatial variation in the intensity of the light beam results from a simultaneous application of a first light beam (LB1) and a second light beam (LB2) to the body, and from interference fringes resulting from an intersection of said first and second light beams.

LTG AlGaAs non-linear optical material and devices fabricated therefrom

Heterostructures having a large lattice mismatch between an upper epilayer and a substrate and a method of forming such structures having a thin intermediate layer are disclosed. The strain due to a lattice mismatch between the intermediate layer and the substrate is partially relieved by the formation of edge type dislocations which are localized and photoelectrically inactive. Growth of the intermediate layer is interrupted before it reaches the thickness at which the left over strain is relieved by 60 degree type threading dislocations. The upper epilayer is then grown in an unstrained and defect-free condition upon the intermediate layer where the unstrained lattice constant of the epilayer is about the same as the partially relieved strain lattice constant or the intermediate layer. An unstrained defect-free epilayer of InGaAs has been grown on a GaAs substrate with an intermediate layer 3-10 nm in thickness of InAs. Other large mismatch systems are disclosed, including, GaAs on Si with an intermediate layer of GaInAs.

Strained defect-free epitaxial mismatched heterostructures and method of

A semiconductor device where an emitter material composition and doping profile produces an electron gas in a base adjacent a band offset heterojunction interface, the electrons in the electron gas in the base are confined under bias by a low barrier and the ballistic carriers have their kinetic energy controlled to prevent intervalley scattering by an electrostatic barrier that under influence of bias provides an essentially level conduction band in the portion of the base adjacent the collector.

Semiconductor ballistic electron velocity control structure

Control of the Fermi level pinning problem and the production of flat band surface performance in compound semiconductors is achieved by providing a cationic oxide free of anionic species on the surface of the semiconductor for flat band performance and with a localized inclusion of some anionic species for barrier performance so that oxide and metal work function responsiveness is available in structure and performance in MOSFET, MESFET and different work function metal FET structures. A cationic gallium oxide is produced on GaAs by oxide growth during illumination and while being rinsed with oxygenated water. The oxidation is used to produce both anionic and cationic species while the rinsing process selectively removes all the anionic species.

Compound semiconductor interface control using cationic ingredient oxide

Submicron structure fabrication is accomplished by providing vapor chemical erosion of a compound crystal by suppressing the more volatile elements so that the less volatile element is provided with an anti-agglomeration and erosion rate limiting capability which can be followed by subsequent regrowth in the same environment. The erosion is sensitive to crystallographic orientation.

Submicron dimension compound semiconductor fabrication using thermal

Growing community of inventors

Putnam Valley, NY, United States of America

Peter Daniel Kirchner