公开(公告)号：WO2019162779A1

公开(公告)日：2019-08-29

申请号：PCT/IB2019/050696

申请日：2019-01-29

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION ,  IBM UNITED KINGDOM LIMITED ,  IBM (CHINA) INVESTMENT COMPANY LIMITED

Inventor： REYES, Sergio ,  TWICHELL, Brian, Chase ,  ZHANG, Yijie ,  INGRAM, Samuel, Kipling

IPC: G06F3/06

Abstract: A data storage system includes first and second storage devices accessed via first and second hardware channels, respectively. A storage controller receives a first access request requesting access to first data and a second access request requesting access to second data, where both the first data and the second data are accessed via the first hardware channel. In response to detecting receipt of the first and second access requests within a same service window defining a range of temporal proximity, the storage controller records an association of identifiers of the first data and second data requested by the first and second access requests, respectively. The storage controller thereafter migrates the second data accessed by the second access request from the first storage device to the second storage device based on the recorded association, such that the first and second data are available for access via different hardware channels.

公开(公告)号：WO2019142087A1

公开(公告)日：2019-07-25

申请号：PCT/IB2019/050272

申请日：2019-01-14

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION ,  IBM UNITED KINGDOM LIMITED ,  IBM (CHINA) INVESTMENT COMPANY LIMITED

Inventor： SMITH, Joshua ,  GIFFORD, Stacey ,  KIM, Sung-Cheol ,  WUNSCH, Benjamin

IPC: C12M1/00

Abstract: Microfluid chips that comprise one or more microscale and/or mesoscale condenser arrays, which can facilitate particle purification and/or fractionation, are described herein. In one embodiment, an apparatus can comprise a layer of a microfluidic chip. The layer can comprise an inlet that can receive fluid,an outlet that can output a purified version of the fluid, and a condenser array coupled between and in fluid communication with the inlet and the outlet. The condenser array can comprise a plurality of pillars arranged in a plurality of columns. Also, a pillar gap sized to facilitate a throughput rate of the fluid of greater than or equal to about 1.0 nanoliter per hour can be located between a first pillar of the plurality of pillars in a first column of the plurality of columns and a second pillar of the plurality of pillars in the first column.

公开(公告)号：WO2019138289A1

公开(公告)日：2019-07-18

申请号：PCT/IB2018/060736

申请日：2018-12-31

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION ,  IBM UNITED KINGDOM LIMITED ,  IBM (CHINA) INVESTMENT COMPANY LIMITED

Inventor： BACARELLA, David ,  BARNEBEE, James ,  LAWRENCE, Nicholas ,  PATEL, Sumit

IPC: G06F17/28

Abstract: A system, computer program product, and method are provided to automate a framework for knowledge graph based persistence of data, and to resolve temporal changes and uncertainties in the knowledge graph. Natural language understanding, together with one or more machine learning models (MLMs), is used to extract data from unstructured information, including entities and entity relationships. The extracted data is populated into a knowledge graph. As the KG is subject to change, the KG is used to create new and retrain existing machine learning models (MLMs). Weighting is applied to the populated data in the form of veracity value. Blockchain technology is applied to the populated data to ensure reliability of the data and to provide auditability to assess changes to the data.

公开(公告)号：WO2019123112A1

公开(公告)日：2019-06-27

申请号：PCT/IB2018/059903

申请日：2018-12-12

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION ,  IBM UNITED KINGDOM LIMITED ,  IBM (CHINA) INVESTMENT COMPANY LIMITED

Inventor： LIN, Chen ,  KALIA, Anup ,  XIAO, Jin ,  VUKOVIC, Maja ,  ANEROUSIS, Nikolaos

IPC: G06F16/36

Abstract: Techniques for generating domain and client-specific application program interface recommendations are provided. A computer-implemented method comprises modifying, by a device operatively coupled to a processor, a description of a client application program interface by removing text data associated with the description of the client application program interface, resulting in a modified description of the client application program interface. The computer-implemented method can further comprise analyzing, by the device, a performance associated with the client application program interface to generate an ontology based on a semantic similarity between the modified description of the client application program interface and one or more previous descriptions of one or more previous client application program interfaces.

公开(公告)号：WO2019123110A1

公开(公告)日：2019-06-27

申请号：PCT/IB2018/059901

申请日：2018-12-12

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION ,  IBM UNITED KINGDOM LIMITED ,  IBM (CHINA) INVESTMENT COMPANY LIMITED

Inventor： SYEDA-MAHMOOD, Tanveer ,  LU, Allen ,  DEHGHAN MARVAST, Ehsan

IPC: G06F17/00

Abstract: Mechanisms are provided to implement an automated echocardiograph measurement extraction system. The automated echocardiograph measurement extraction system receives medical imaging data comprising one or more medical images and inputs the one or more medical images into a deep learning network. The deep learning network automatically processes the one or more medical images to generate an extracted echocardiograph measurement vector output comprising one or more values for echocardiograph measurements extracted from the one or more medical images. The deep learning network outputs the extracted echocardiograph measurement vector output to a medical image viewer.

公开(公告)号：WO2019123068A1

公开(公告)日：2019-06-27

申请号：PCT/IB2018/059607

申请日：2018-12-04

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION ,  IBM UNITED KINGDOM LIMITED ,  IBM (CHINA) INVESTMENT COMPANY LIMITED

Inventor： CIDECIYAN, Roy ,  FURRER, Simeon ,  LANTZ, Mark, Alfred

IPC: G11B20/18

Abstract: A method includes writing a data set to a sequential access medium. The method also includes reading the data set immediately after being written to the sequential access medium in a read-while-write process to identify one or more faulty encoded data blocks, each of the one or more faulty encoded data blocks including at least one faulty codeword having symbols at least 10 bits in size. Moreover, the method includes rewriting a first of the one or more faulty encoded data blocks within a first encoded data block set to a particular logical track in the rewrite area of the sequential access medium selected from a predetermined subset of logical tracks. The predetermined subset of logical tracks includes D1+D2+1 logical tracks. Only one faulty encoded data block from a particular sub data set is rewritten in a single encoded data block set in the rewrite area.

公开(公告)号：WO2019116152A1

公开(公告)日：2019-06-20

申请号：PCT/IB2018/059558

申请日：2018-12-03

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION ,  IBM UNITED KINGDOM LIMITED ,  IBM (CHINA) INVESTMENT COMPANY LIMITED

Inventor： LI, Juntao ,  CHENG, Kangguo ,  JIANG, Liying ,  GAUDIELLO, John, Gerard

IPC: H01L21/84

Abstract: A method of forming a logic device and a power device on a substrate is provided. The method includes forming a first vertical fin on a first region of the substrate and a second vertical fin on a second region of the substrate, wherein an isolation region separates the first region from the second region, forming a dielectric under-layer segment on the second vertical fin on the second region, and forming a first gate structure on the dielectric under- layer segment and second vertical fin on the second region.

公开(公告)号：WO2019116137A1

公开(公告)日：2019-06-20

申请号：PCT/IB2018/059453

申请日：2018-11-29

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION ,  IBM UNITED KINGDOM LIMITED ,  IBM (CHINA) INVESTMENT COMPANY LIMITED

Inventor： GKOULALAS-DIVANIS, Aris

IPC: G06K9/00

CPC classification number: G06F21/6254 ,  G06F21/577 ,  G06F2221/033

Abstract: A system for de-identifying data determines one or more identifiers that identify an entity of a dataset. One or more data de-identification processes are identified and associated with the determined one or more identifiers. Each data de-identification process is associated with one or more sets of configuration options indicating information to preserve in the dataset. The identified data de-identification processes are executed on the dataset in accordance with the associated sets of configuration options to generate datasets with varying preserved information. The generated datasets are evaluated for privacy vulnerabilities and a data de-identification process and an associated set of configuration options are selected based on the evaluation. The selected data de-identification process is executed on the dataset according to the associated set of configuration options to produce a resulting de-identified data set. Embodiments include a method and computer program product for de-identifying data in substantially the same manner described above.

公开(公告)号：WO2019111118A1

公开(公告)日：2019-06-13

申请号：PCT/IB2018/059516

申请日：2018-11-30

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION ,  IBM UNITED KINGDOM LIMITED ,  IBM (CHINA) INVESTMENT COMPANY LIMITED

Inventor： CHEN, Chia-Yu ,  AGRAWAL, Ankur ,  BRAND, Daniel ,  GOPALAKRISHNAN, Kailash ,  CHOI, Jungwook

IPC: G06N3/08

Abstract: Embodiments of the present invention provide a computer-implemented method for adaptive residual gradient compression for training of a deep learning neural network (DNN). The method includes obtaining, by a first learner, a current gradient vector for a neural network layer of the DNN, in which the current gradient vector includes gradient weights of parameters of the neural network layer that are calculated from a mini-batch of training data. A current residue vector is generated that includes residual gradient weights for the mini-batch. A compressed current residue vector is generated based on dividing the residual gradient weights of the current residue vector into a plurality of bins of a uniform size and quantizing a subset of the residual gradient weights of one or more bins of the plurality of bins. The compressed current residue vector is then transmitted to a second learner of the plurality of learners or to a parameter server.

公开(公告)号：WO2019106516A1

公开(公告)日：2019-06-06

申请号：PCT/IB2018/059296

申请日：2018-11-26

Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION ,  IBM UNITED KINGDOM LIMITED ,  IBM (CHINA) INVESTMENT COMPANY LIMITED

Inventor： CAO, Qing ,  TANG, Jianshi ,  LI, Ning

IPC: H01L27/146

CPC classification number: H01L29/1606 ,  G01J3/2823 ,  G01J3/447 ,  G01J2003/2826 ,  G01N21/255 ,  G01N21/314 ,  H01L27/14601 ,  H01L27/14665 ,  H04N5/332

Abstract: A vertically integrated multispectral imaging sensor includes a first metal contact layer on a substrate, an SiO 2 layer on the first metal contact layer with a first detector element embedded in a hole therein, a first graphene layer that covers the first detector element, a second metal contact layer on the SiO 2 layer on one side of the first graphene, an AlO 3 layer on the SiO 2 layer, in which a second detector element is embedded in a hole over the first graphene layer, a second graphene layer on the second detector element, and a third metal contact layer on the AlO 3 layer adjacent to the second graphene layer. The first detector material is sensitive to a different wavelength band of the electromagnetic spectrum than the second detector material.