中科微点-全球专利检索

  1. Login
  2. Sign Up

Search Results

21 records (took 0.033 seconds)

    Training a cognitive system on partial correctness
    1.
    发明授权
    Training a cognitive system on partial correctness 有权

    公开(公告)号：US11093859B2

    公开(公告)日：2021-08-17

    申请号：US15845864

    申请日：2017-12-18

    Applicant: International Business Machines Corporation

    Inventor： Kelley L. Anders ,  Paul K. Bullis ,  Geoffrey M. Hambrick

    IPC: G06N20/00 ,  G06N5/04 ,  G06F40/30 ,  G06F40/205

    Abstract: A set of components is computed from performing NLP on a question in an input. An actual answer is computed corresponding to the question by a cognitive system. the actual answer corresponds to an actual subset of the set of components, and an expected answer corresponds to an expected subset of the subset of components. The actual answer is mapped to an actual category in a hierarchy of answer categories. A distance between the expected answer and the actual answer is computed where the distance is a function of a path in the hierarchy from the actual category to the expected category, and a degree of correctness of the actual answer is another function of the distance. A self-learning operation in the cognitive system causes a revised actual answer on the question being at a shorter distance from the expected answer.

    TRAINING A COGNITIVE SYSTEM ON PARTIAL CORRECTNESS
    2.
    发明申请
    TRAINING A COGNITIVE SYSTEM ON PARTIAL CORRECTNESS 审中-公开

    公开(公告)号：US20190130311A1

    公开(公告)日：2019-05-02

    申请号：US15845864

    申请日：2017-12-18

    Applicant: International Business Machines Corporation

    Inventor： Kelley L. Anders ,  Paul K. Bullis ,  Geoffrey M. Hambrick

    IPC: G06N99/00 ,  G06N5/04 ,  G06F17/27

    Abstract: A set of components is computed from performing NLP on a question in an input. An actual answer is computed corresponding to the question by a cognitive system. the actual answer corresponds to an actual subset of the set of components, and an expected answer corresponds to an expected subset of the subset of components. The actual answer is mapped to an actual category in a hierarchy of answer categories. A distance between the expected answer and the actual answer is computed where the distance is a function of a path in the hierarchy from the actual category to the expected category, and a degree of correctness of the actual answer is another function of the distance. A self-learning operation in the cognitive system causes a revised actual answer on the question being at a shorter distance from the expected answer.

    SEARCH JOURNALING FOR OPERATIONS ANALYSIS
    4.
    发明申请
    SEARCH JOURNALING FOR OPERATIONS ANALYSIS 审中-公开

    公开(公告)号：US20170364566A1

    公开(公告)日：2017-12-21

    申请号：US15184641

    申请日：2016-06-16

    Applicant: International Business Machines Corporation

    Inventor： Paul K. Bullis ,  Geoffrey M. Hambrick

    IPC: G06F17/30

    CPC classification number: G06F16/248 ,  G06F16/1805 ,  G06F16/2246

    Abstract: A session initiation with an information system is detected, the session including a search for a data item in the information system. In an application executing using a processor and a memory, a node is constructed in a search tree, the node representing a search-related step occurring in the session. The node is configured with at least a search term used in the search and a context within which the search is performed using the search term. The node is positioned relative to a second node in the search tree, where the second node corresponds to a second search that was performed prior to the search in the session. The search tree is output as a part of a search journal corresponding to the session.

    ADAPTIVE TASK ASSIGNMENT TO IMPROVE SKILL GROWTH AND BALANCE SERVICE LEVEL AGREEMENTS
    6.
    发明申请
    ADAPTIVE TASK ASSIGNMENT TO IMPROVE SKILL GROWTH AND BALANCE SERVICE LEVEL AGREEMENTS 审中-公开

    公开(公告)号：US20160140479A1

    公开(公告)日：2016-05-19

    申请号：US14726473

    申请日：2015-05-30

    Applicant: International Business Machines Corporation

    Inventor： Geoffrey M. Hambrick ,  Govinda Kamtamneni

    IPC: G06Q10/06

    CPC classification number: G06Q10/063112 ,  G06Q10/063118

    Abstract: Methods and systems may provide for computing the complexities of tasks in a business process management system based on business and process data, and previous durations taken to compute similar tasks. The system and method may also categorize the skill level of one or more users based on user history and other metadata, and use the skill level and task complexity to correlate a task with one or more users that are suitable to complete the task within a related service level agreement. Tasks may be adaptively assigned to users in order to ensure that the tasks are completed as required by the SLA, assigned to less skilled users to improve their skills, and/or dynamically reassigned to ensure that the requirements of the SLA are met. Additionally, the system and method may also be used to manually configure complexities of a task and tag users based on skill level.

    Training a cognitive system on partial correctness
    8.
    发明授权
    Training a cognitive system on partial correctness 有权

    公开(公告)号：US11093858B2

    公开(公告)日：2021-08-17

    申请号：US15797608

    申请日：2017-10-30

    Applicant: International Business Machines Corporation

    Inventor： Kelley L. Anders ,  Paul K. Bullis ,  Geoffrey M. Hambrick

    IPC: G06N20/00 ,  G06N5/04 ,  G06F40/30 ,  G06F40/205

    Abstract: A set of components is computed from performing NLP on a question in an input. An actual answer is computed corresponding to the question by a cognitive system. the actual answer corresponds to an actual subset of the set of components, and an expected answer corresponds to an expected subset of the subset of components. The actual answer is mapped to an actual category in a hierarchy of answer categories. A distance between the expected answer and the actual answer is computed where the distance is a function of a path in the hierarchy from the actual category to the expected category, and a degree of correctness of the actual answer is another function of the distance. A self-learning operation in the cognitive system causes a revised actual answer on the question being at a shorter distance from the expected answer.

    SELF-LEARNING CONTEXTUAL MODALITY SELECTION FOR COGNITIVE SOLUTION DELIVERY
    9.
    发明申请
    SELF-LEARNING CONTEXTUAL MODALITY SELECTION FOR COGNITIVE SOLUTION DELIVERY 审中-公开

    公开(公告)号：US20190138945A1

    公开(公告)日：2019-05-09

    申请号：US15851315

    申请日：2017-12-21

    Applicant: International Business Machines Corporation

    Inventor： Peter C. Bahrs ,  Paul K. Bullis ,  Geoffrey M. Hambrick

    IPC: G06N99/00 ,  G06N5/04

    Abstract: A problem context is computed from an input at an application. The problem context includes a set of problem factors, the input including a problem to be solved using a cognitive system. A user context is computed from the input at the application, the user context including a set of user factors. A type of media is determined corresponding to a complexity of a cognitive solution received from the cognitive system, where the cognitive solution is in response to the problem. Using a problem factor from the set of problem factors, using a user factor in the set of user factors, and the complexity, a mode of communication is determined. A communication apparatus is adjusted to cause a data communication to occur and deliver the cognitive solution in the type of media using the mode of communication.

    Managing reuse of assets in a workflow management system
    10.
    发明授权
    Managing reuse of assets in a workflow management system 有权

    公开(公告)号：US10127024B2

    公开(公告)日：2018-11-13

    申请号：US15191127

    申请日：2016-06-23

    Applicant: International Business Machines Corporation

    Inventor： Paul K. Bullis ,  Geoffrey M. Hambrick

    IPC: G06F9/44 ,  G06F8/36 ,  G06F8/70

    Abstract: A category is determined of a task being planned at a given time. A collection of assets is identified, the collection having been used in a similar task at a past time. An experience report associated with an asset in the collection is extracted into a set of encapsulated experience reports, where the experience report was generated from the similar task at the past time. A degree of usability is computed corresponding to the collection, the collection being usable in the category of the task up to the degree of usability. According to the degree of usability of the collection, the collection is positioned in a sorted list of reusable assets. A reusable asset is selected from the sorted list for use in the task.

Patent Agency Ranking