知搜-全球专利检索

  1. Login
  2. Sign Up

Search Results

30 records (took 0.021 seconds)

    Scalable ground truth disambiguation
    11.
    发明授权
    Scalable ground truth disambiguation 有权

    公开(公告)号：US10572826B2

    公开(公告)日：2020-02-25

    申请号：US15490081

    申请日：2017-04-18

    Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION

    Inventor： Faheem Altaf ,  Lisa Seacat Deluca ,  Raghuram Srinivas

    IPC: G06F17/00 ,  G06N20/00

    Abstract: Methods, computer program products, and systems are presented. The methods include, for instance: obtaining an utterance input from a user agent, and collecting context data of the utterance input. A context tag is generated based on the context data, and one or more ground truth having respective utterance semantically identical to the utterance input is selected. Semantical relationship between the context tag and an intent of the selected ground truth is examined and the selected ground truth is updated with the context tag.

    FOOD RECIPE SCORING AND RANKING SYSTEM
    13.
    发明申请
    FOOD RECIPE SCORING AND RANKING SYSTEM 审中-公开

    公开(公告)号：US20160103834A1

    公开(公告)日：2016-04-14

    申请号：US14513592

    申请日：2014-10-14

    Applicant: International Business Machines Corporation

    Inventor： Faheem Altaf ,  Florian Pinel ,  Nan Shao

    IPC: G06F17/30 ,  G06F19/28

    CPC classification number: G06F3/0482 ,  G06Q30/0282 ,  G16H20/60

    Abstract: According to an aspect, food recipes are scored and ranked based on one or more culinary objectives. A general dish specification includes a selected dish type, a selected cuisine type, and one or more culinary objectives. Candidate ingredient types are identified based on recipes of the selected dish type and the selected cuisine type in a recipe database. A number or range of ingredients are computed for each candidate ingredient type based on a frequency of occurrence of ingredients of each candidate ingredient type. A recipe closeness score is computed for a plurality of the recipes of the selected dish type and the selected cuisine type in the recipe database based on the culinary objectives, a bill of ingredients, and parameters for the recipes stored in the database. The recipes are ranked based on the recipe closeness scores and a ranked list of the recipes is provided.

    Superposition detection and correction
    14.
    发明授权
    Superposition detection and correction 有权

    公开(公告)号：US11032435B1

    公开(公告)日：2021-06-08

    申请号：US16744611

    申请日：2020-01-16

    Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION

    Inventor： Igor S. Ramos ,  Andrew J. Lavery ,  Faheem Altaf ,  Paul Hake

    IPC: G06K9/03 ,  H04N1/00 ,  G06F16/23 ,  G06K9/46

    Abstract: A method, computer system, and computer program product for document superposition detection and correction are provided. The embodiment may include receiving a plurality of scanned documents. The embodiment may also include analyzing the plurality of the scanned documents. The embodiment may further include determining if there is a superposition scanning defect within the scanned documents based on the analysis. The embodiment may also include in response to the plurality of the scanned documents containing the superposition scanning defect, identifying a subset of documents within the plurality of scanned documents containing the determined superposition scanning defect. The embodiment may further include transmitting a notification to a user, wherein the notification identifies the subset and that the subset requires a rescan.

    NETWORK MANAGED RULES FOR MACHINE ACCESS
    15.
    发明申请
    NETWORK MANAGED RULES FOR MACHINE ACCESS 有权

    公开(公告)号：US20210090363A1

    公开(公告)日：2021-03-25

    申请号：US16582935

    申请日：2019-09-25

    Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION

    Inventor： Igor S. Ramos ,  Andrew J. Lavery ,  Leucir Marin, JR. ,  Faheem Altaf

    IPC: G07C9/00 ,  G06N20/00 ,  G06F21/42

    Abstract: Provided are techniques for network managed rules for machine access. A request to access a machine is received from an operator. One or more machine access rules are identified for a machine and an operator who will operate the machine for a company in a location, where each of the one or more machine access rules includes variable criteria. Data for the variable criteria of the one or more machine access rules is obtained from one or more of: a machine profile of the machine, an operator profile of the operator, a company profile of the company, and environmental data for the location. The data is applied to the variable criteria of the one or more machine access rules to determine whether to provide the operator with access to the machine at the location. An indication of whether the operator has access to the machine at the location is determined.

    WEBSITE DOMAIN SPECIFIC SEARCH
    17.
    发明申请
    WEBSITE DOMAIN SPECIFIC SEARCH 审中-公开

    公开(公告)号：US20200012700A1

    公开(公告)日：2020-01-09

    申请号：US16574390

    申请日：2019-09-18

    Applicant: INTERNATIONAL BUSINESS MACHINES CORPORATION

    Inventor： Faheem Altaf ,  Lisa Seacat DeLuca ,  Raghuram Srinivas

    IPC: G06F16/9535

    Abstract: A method and system for improving a domain specific search is provided. The method includes executing a user initiated search query and analyzing associated. In response, a search results data set is generated and associated hardware sensor devices detect how a user interacts with specific search facets of the search results data set. The search results data set is refined and attributes of the specific search facets are determined. Unstructured data associated with items described within the search results data set is updated.

    COGNITIVE RETRIEVE AND RANK SEARCH IMPROVEMENTS USING NATURAL LANGUAGE FOR PRODUCT ATTRIBUTES
    19.
    发明申请
    COGNITIVE RETRIEVE AND RANK SEARCH IMPROVEMENTS USING NATURAL LANGUAGE FOR PRODUCT ATTRIBUTES 审中-公开

    公开(公告)号：US20190213277A1

    公开(公告)日：2019-07-11

    申请号：US15868558

    申请日：2018-01-11

    Applicant: International Business Machines Corporation

    Inventor： Lisa Seacat DeLuca ,  Raghuram Srinivas ,  Faheem Altaf

    IPC: G06F17/30 ,  G06Q30/06

    Abstract: A method, computer system, and computer readable program product for generating text for a search corpus. In an embodiment, the method comprises analyzing structured data associated with an entity; and breaking the structured data into multiple unstructured natural language pair attributes such that natural language search terms have an increased level of matching with attributes associated with the entity, including filtering the structured data to identify a filtered group of attributes, and using the filtered group of attributes to form the natural language pair attributes. The method further comprises saving the formed natural language pairs in a search corpus. In an embodiment, the method further comprises generating a cluster of the attributes based on semantic similarities; and the using the filtered gropo of attributes to form the natural language pair attributes includes creating separate blocks of text that respectively describe the attributes included in the cluster of attributes.

Patent Agency Ranking