中科微点-全球专利检索

  1. Login
  2. Sign Up

Search Results

38 records (took 0.031 seconds)

    TRANSFER KNOWLEDGE FROM AUXILIARY DATA FOR MORE INCLUSIVE MACHINE LEARNING MODELS
    21.
    发明公开
    TRANSFER KNOWLEDGE FROM AUXILIARY DATA FOR MORE INCLUSIVE MACHINE LEARNING MODELS 审中-公开

    公开(公告)号：US20240104422A1

    公开(公告)日：2024-03-28

    申请号：US17935747

    申请日：2022-09-27

    Applicant: AT&T Intellectual Property I, L.P.

    Inventor： Zhengyi Zhou ,  Cheryl Brooks ,  Aritra Guha ,  Yaron Kanza ,  Balachander Krishnamurthy

    IPC: G06N20/00

    CPC classification number: G06N20/00

    Abstract: Transfer knowledge from auxiliary data for more inclusive machine learning models is provided. A method can include generating a common feature space comprising first data features, wherein the first data features are present in training data used to train a first machine learning model, and wherein the first data features are present in auxiliary data that are independent of the training data; generating a combined learned feature representation, the combined learned feature representation being representative of the first data features of the common feature space and second data features that are unique to the training data; and training a second machine learning model based on the combined learned feature representation.

    Bias scoring of machine learning project data
    22.
    发明授权
    Bias scoring of machine learning project data 有权

    公开(公告)号：US11620542B2

    公开(公告)日：2023-04-04

    申请号：US16704965

    申请日：2019-12-05

    Applicant: AT&T Intellectual Property I, L.P.

    Inventor： Emily Dodwell ,  Balachander Krishnamurthy ,  Rajat Malik ,  Ritwik Mitra

    IPC: G06N5/04 ,  G06F16/28 ,  G06N20/00

    Abstract: Aspects of the subject disclosure may include, for example, system and apparatus that enable operations that may include receiving, by a processing system, project data defining a proposed machine learning (ML) project of an entity and storing the project data in a project database with other project data for other projects. The operations may further include extracting extracted features of the proposed project and, based on the extracted features, determining a clustering assignment for the proposed project. Determining the clustering assignment may comprise comparing information about the proposed project including the extracted features with information about the other projects and assigning the proposed project to a cluster including one or more projects having similar bias characteristics as the proposed project. The operations may further include determining a risk of potential bias for the proposed project and, based on the risk of bias, recommending a corrective action to reduce the risk of bias. Machine learning models may be used for project clustering and bias score determination and may be readily updated as new ML projects are evaluated. Other embodiments are disclosed.

    Methods, systems, and devices for detecting and mitigating potential bias
    23.
    发明授权
    Methods, systems, and devices for detecting and mitigating potential bias 有权

    公开(公告)号：US11586950B2

    公开(公告)日：2023-02-21

    申请号：US16705520

    申请日：2019-12-06

    Applicant: AT&T Intellectual Property I, L.P.

    Inventor： Emily Dodwell ,  Balachander Krishnamurthy ,  Ritwik Mitra

    IPC: G06F15/16 ,  G06N5/04 ,  G06Q30/0241 ,  G06N20/00

    Abstract: Aspects of the disclosure include, for example, obtaining input data. Further embodiments include a determination of a fast path prediction for a first time period according to the input data based on a fast path model. Embodiments include providing instructions to deliver information to a user device according to the fast path prediction. Additional embodiments include obtaining additional input data. Embodiments include a determination of a slow path prediction for the first time period according to the input data and the additional input data based on a slow path model, retraining the fast path model according to the input data and the fast path prediction, and training the slow path model according to the slow path prediction. Embodiments include a determination of a fast path negative impact metric and determination of a slow path negative impact metric. Other embodiments are disclosed.

    METHODS, SYSTEMS, AND DEVICES FOR SELF-CERTIFICATION OF BIAS ABSENCE
    24.
    发明申请
    METHODS, SYSTEMS, AND DEVICES FOR SELF-CERTIFICATION OF BIAS ABSENCE 有权

    公开(公告)号：US20220005077A1

    公开(公告)日：2022-01-06

    申请号：US16919457

    申请日：2020-07-02

    Applicant: AT&T Intellectual Property I, L.P.

    Inventor： Balachander Krishnamurthy ,  Subhabrata Majumdar ,  Ritwik Mitra ,  David Poole

    IPC: G06Q30/02 ,  H04W8/18 ,  H04W16/24 ,  H04W88/14 ,  G06N20/00

    Abstract: Aspects of the subject disclosure may include, for example, embodiments receiving a notification of actions, determining a potential bias metric for the actions in response to analyzing the actions using a machine learning application, determining the potential bias metric for the actions is above a potential bias threshold for the actions, and adjusting the actions to mitigate potential bias in the actions according to the potential bias metric being above the potential bias threshold using the machine learning application. Further embodiments can include determining a potential bias metric for the adjusted actions in response to analyzing the adjusted actions using the machine learning application, determining the potential bias metric for the adjusted actions is below the potential bias threshold for the actions, and providing a notification that indicates to implement the adjusted actions. Other embodiments are disclosed.

    METHODS, SYSTEMS, AND DEVICES FOR DETECTING AND MITIGATING POTENTIAL BIAS
    25.
    发明申请
    METHODS, SYSTEMS, AND DEVICES FOR DETECTING AND MITIGATING POTENTIAL BIAS 有权

    公开(公告)号：US20210174223A1

    公开(公告)日：2021-06-10

    申请号：US16705520

    申请日：2019-12-06

    Applicant: AT&T Intellectual Property I, L.P.

    Inventor： Emily Dodwell ,  Balachander Krishnamurthy ,  Ritwik Mitra

    IPC: G06N5/04 ,  G06N20/00 ,  G06Q30/02

    Abstract: Aspects of the disclosure include, for example, obtaining input data. Further embodiments include a determination of a fast path prediction for a first time period according to the input data based on a fast path model. Embodiments include providing instructions to deliver information to a user device according to the fast path prediction. Additional embodiments include obtaining additional input data. Embodiments include a determination of a slow path prediction for the first time period according to the input data and the additional input data based on a slow path model, retraining the fast path model according to the input data and the fast path prediction, and training the slow path model according to the slow path prediction. Embodiments include a determination of a fast path negative impact metric and determination of a slow path negative impact metric. Other embodiments are disclosed.

    Systems and Methods for Rule-Based Anomaly Detection on IP Network Flow
    29.
    发明申请
    Systems and Methods for Rule-Based Anomaly Detection on IP Network Flow 有权
    基于规则的IP网络异常检测系统与方法

    公开(公告)号：US20160105462A1

    公开(公告)日：2016-04-14

    申请号：US14969591

    申请日：2015-12-15

    Applicant: AT&T Intellectual Property I, L.P.

    Inventor： Nicholas Duffield ,  Patrick Haffner ,  Balachander Krishnamurthy ,  Haakon Andreas Ringberg

    IPC: H04L29/06 ,  G06N5/02 ,  H04L12/721 ,  H04L12/26 ,  H04L12/24

    CPC classification number: H04L63/20 ,  G06F21/552 ,  G06N5/025 ,  H04L41/00 ,  H04L41/0604 ,  H04L41/16 ,  H04L43/026 ,  H04L43/028 ,  H04L43/16 ,  H04L45/28 ,  H04L45/38 ,  H04L47/10 ,  H04L47/24 ,  H04L63/14 ,  H04L63/1416 ,  H04L63/1425 ,  H04L63/30

    Abstract: A system to detect anomalies in internet protocol (IP) flows uses a set of machine-learning (ML) rules that can be applied in real time at the IP flow level. A communication network has a large number of routers equipped with flow monitoring capability. A flow collector collects flow data from the routers throughout the communication network and provides them to a flow classifier. At the same time, a limited number of locations in the network monitor data packets and generate alerts based on packet data properties. The packet alerts and the flow data are provided to a machine learning system that detects correlations between the packet-based alerts and the flow data to thereby generate a series of flow-level alerts. These rules are provided to the flow time classifier. Over time, the new packet alerts and flow data are used to provide updated rules generated by the machine learning system.

    Abstract translation: 检测互联网协议（IP）流中的异常的系统使用一组机器学习（ML）规则，可以在IP流级别实时应用。 通信网络具有大量具有流量监控功能的路由器。 集流器在通信网络中收集来自路由器的流数据，并将其提供给流分类器。 同时，网络中有限数量的位置监视数据包，并根据数据包数据属性生成警报。 分组警报和流数据被提供给机器学习系统，其检测基于分组的警报和流数据之间的相关性，从而生成一系列流级别警报。 这些规则提供给流时间分类器。 随着时间的推移，新的数据包警报和流数据用于提供机器学习系统生成的更新规则。

Patent Agency Ranking