利索-全球专利检索

  1. 登录
  2. 注册

搜索结果

7 条记录 (耗时 0.037 秒)

    METHODS AND SYSTEMS FOR CROSS-DOMAIN FEW-SHOT CLASSIFICATION
    4.
    发明申请
    METHODS AND SYSTEMS FOR CROSS-DOMAIN FEW-SHOT CLASSIFICATION 有权

    公开(公告)号：US20220300823A1

    公开(公告)日：2022-09-22

    申请号：US17204670

    申请日：2021-03-17

    申请人： Hanwen LIANG ,  Peng DAI ,  Qiong ZHANG ,  Juwei LU

    发明人： Hanwen LIANG ,  Peng DAI ,  Qiong ZHANG ,  Juwei LU

    IPC分类号： G06N3/08 ,  G06N3/04

    摘要： Methods, systems, and media for training deep neural networks for cross-domain few-shot classification are described. The methods comprise an encoder and a decoder of a deep neural network. The training of the autoencoder comprises two training stages. For each iteration in the first training stage, a batch of data samples from the source dataset are sampled and fed to the encoder to generate a plurality of source feature maps, then determining a first training stage loss, which updates the autoencoder's parameters. For each iteration in the second training stage, the novel dataset is split into a support set and a query set. The support set is fed to the encoder to determine a prototype for each class label. The query set is also fed to the encoder to calculate a query set metric classification loss. The query set metric classification loss updates the autoencoder's parameters.

    PROBABILITY-BASED REGENERATOR SITE ANALYSIS
    5.
    发明申请
    PROBABILITY-BASED REGENERATOR SITE ANALYSIS 有权
    基于概率的再生器站点分析

    公开(公告)号：US20140052419A1

    公开(公告)日：2014-02-20

    申请号：US13585663

    申请日：2012-08-14

    申请人： Xi WANG ,  Qiong ZHANG ,  Paparao PALACHARLA ,  Motoyoshi SEKIYA

    发明人： Xi WANG ,  Qiong ZHANG ,  Paparao PALACHARLA ,  Motoyoshi SEKIYA

    IPC分类号： G06G7/62 ,  G06F17/10

    CPC分类号： H04Q11/00 ,  H04J14/00 ,  H04L41/142 ,  H04L41/145 ,  H04Q2213/13164 ,  H04Q2213/13335

    摘要： An optical network analysis tool includes a computer-readable storage medium having computer-readable instructions stored thereon. The computer-readable instructions are executable by a computing device to perform operations. The operations include generating a simulated network that models an optical network. The simulated network includes regenerator candidate sites. The operations may also include conducting an analysis of the optical network. The analysis includes introducing a multiple signals transmitted between source/destination pairs and recording a number of times each of the regenerator candidate sites are selected as a regenerator site while applying each of a set of data traffic conditions in the simulated network. The operations may also include statistically analyzing the number of times each of the regenerator candidate sites is selected to generate statistically analyzed information and presenting the statistically analyzed information.

    摘要翻译： 光网络分析工具包括其上存储有计算机可读指令的计算机可读存储介质。 计算机可读指令可由计算设备执行以执行操作。 这些操作包括生成模拟光网络的模拟网络。 模拟网络包括再生器候选站点。 这些操作还可以包括进行光网络的分析。 分析包括引入在源/目的地对之间传输的多个信号，并且在将模拟网络中的一组数据业务条件应用中的每一个的同时，将每个再生器候选站点选择为多个次数作为再生站站点进行记录。 操作还可以包括统计分析每个再生器候选位点被选择以产生统计分析的信息并呈现统计分析的信息的次数。

    LARGE-SCALE ITEM AFFINITY DETERMINATION USING A MAP REDUCE PLATFORM
    7.
    发明申请
    LARGE-SCALE ITEM AFFINITY DETERMINATION USING A MAP REDUCE PLATFORM 审中-公开
    使用地图减少平台进行大规模的物品确定

    公开(公告)号：US20100205075A1

    公开(公告)日：2010-08-12

    申请号：US12369160

    申请日：2009-02-11

    申请人： Qiong ZHANG

    发明人： Qiong ZHANG

    IPC分类号： G06N5/02 ,  G06Q10/00 ,  G06F17/30 ,  G06F17/40 ,  G06F7/06

    CPC分类号： G06Q10/04 ,  G06Q40/12

    摘要： Pair-wise item affinity is based on transaction records. Each transaction record includes an indication of a bucket and an indication of an item transacted corresponding to that bucket. The method comprises a Phase 1 bucket filtering, Phase 2 item count, Phase 3 bucket materialization and Phase 4 pair count and affinity lift/calculation. The phases are ideally suited to be carried out by a computing system in a map-reduce configuration.

    摘要翻译： 成对项目关联性基于事务记录。 每个交易记录包括桶的指示和对应于该桶的事务的指示。 该方法包括阶段1桶过滤，阶段2项目计数，阶段3桶实现和阶段4对计数和亲和力提升/计算。 这些阶段非常适合由地图缩减配置中的计算系统执行。