EFFICIENT INFERENCING WITH FAST POINTWISE CONVOLUTION

    公开(公告)号:US20210081765A1

    公开(公告)日:2021-03-18

    申请号:US16571760

    申请日:2019-09-16

    摘要: Embodiments described herein relate to a method, comprising: receiving input data at a convolutional neural network (CNN) model; generating a factorized computation network comprising a plurality of connections between a first layer of the CNN model and a second layer of the CNN model, wherein: the factorized computation network comprises N inputs, the factorized computation network comprises M outputs, and the factorized computation network comprises at least one path from every input of the N inputs to every output of the M outputs; setting a connection weight for a plurality of connections in the factorized computation network to 1 so that a weight density for the factorized computation network is

    GENERATION OF POLAR CODES WITH A VARIABLE BLOCK LENGTH UTILIZING PUNCTURING

    公开(公告)号:US20200274644A1

    公开(公告)日:2020-08-27

    申请号:US16874264

    申请日:2020-05-14

    IPC分类号: H04L1/00 H03M13/13 H03M13/00

    摘要: Polar codes may be generated with a variable block length utilizing puncturing. Some puncturing schemes consider punctured bits as unknown bits, and set the log likelihood ratio (LLR) for those bits to zero; while other puncturing schemes consider punctured bits as known bits, and set the LLR for those bits to infinity. Each of these puncturing schemes has been observed to provide benefits over the other under different circumstances, especially corresponding to different coding rates or different signal to noise ratio (SNR). According to aspects of the present disclosure, both puncturing schemes are compared, and the puncturing scheme resulting in the better performance is utilized for transmission.

    LINEAR PRECODING IN FULL-DIMENSIONAL MIMO SYSTEMS AND DYNAMIC VERTICAL SECTORIZATION

    公开(公告)号:US20200014437A1

    公开(公告)日:2020-01-09

    申请号:US16573304

    申请日:2019-09-17

    摘要: Certain aspects of the present disclosure provide methods and apparatus for linear precoding in full-dimensional MIMO (FD-MIMO) systems. According to aspects, an eNB may compress a larger number of antenna elements to a smaller number of antenna ports. The eNB may use a port precoding matrix to transmit reference signals to a UE, receive feedback regarding CSI based on the reference signals, and transmit data to the UE, based on a mapping of multiple data layers and mapping of antenna ports to the physical antenna elements. Further, aspects include performing elevation beamforming by dynamically forming one or more vertical sectors based on UE feedback in the elevation domain.

    TECHNIQUES FOR CHANNEL STATE INFORMATION ACQUISITION IN NEW RADIO TECHNOLOGY

    公开(公告)号:US20190190581A1

    公开(公告)日:2019-06-20

    申请号:US16324710

    申请日:2017-08-11

    IPC分类号: H04B7/06 H04L1/00

    摘要: Certain aspects of the present disclosure provide techniques for channel state information acquisition in wireless communication systems operating according to new radio (NR) technologies. An exemplary method that may be performed by a user equipment (UE) generally includes receiving a first beamformed channel state information reference signal (CSI-RS), determining, based on the first beamformed CSI-RS and the beacon reference signal, information regarding an adjustment, preferred by the UE, of a beam-former to be used for a forthcoming second beamformed CSI-RS, transmitting a beamformer adjustment indicator (BAI), periodically, semi-persistently, or aperiodically, that indicates the preferred adjustment, and reporting channel state information feedback (CSF), based on the first beamformed CSI-RS.

    UE-RS-BASED OPEN-LOOP AND SEMI-OPEN-LOOP MIMO

    公开(公告)号:US20190089429A1

    公开(公告)日:2019-03-21

    申请号:US16078526

    申请日:2017-03-22

    摘要: Design of precoding and feedback for user equipment (UE)-specific reference signals (UE-RS)-based open-loop and semi-open-loop multiple input, multiple output (MIMO) systems is discussed. Aspects of the present disclosure provide for sub-resource block (RB) random precoding that allows for greater diversity gain in a lower bandwidth. In addition, the recoding may be performed using resource element (RE)-level layer shifting that provides for a number of precoders to be assigned to a number of layers for every such continuous subcarrier. As such, two codewords may experience the same effective channel quality with channel quality indicators (CQI) being averaged across all of the layers.

    TECHNIQUES FOR FLEXIBLE DUPLEXING
    56.
    发明申请

    公开(公告)号:US20180192400A1

    公开(公告)日:2018-07-05

    申请号:US15738990

    申请日:2016-07-01

    摘要: The present disclosure, for example, generally relates to wireless communication systems, and more particularly to techniques for flexible duplexing in such systems. For example, a technique for flexible duplexing provides a reference configuration that determines characteristics for paired frequency division duplexing (FDD) bands when the uplink band is temporarily reconfigured for time division duplexing (TDD) use. A user equipment (UE) may use the reference configuration to determine hybrid automatic repeat request (HARQ) timing, schedule uplink transmissions, manage a soft buffer, and determine signaling formats. In an aspect, the UE may receive a reconfiguration message indicating a change for an FDD uplink band to a temporary TDD band. The UE may then determine a reference configuration for a pair of FDD bands including an FDD downlink band and the temporary TDD band, the reference configuration indicating a pattern of sub-frames associated with the pair of FDD bands.

    APERIODIC CQI REPORTING FOR LTE-TDD EIMTA SYSTEM
    57.
    发明申请
    APERIODIC CQI REPORTING FOR LTE-TDD EIMTA SYSTEM 有权
    LTE-TDD EIMTA系统的APERIODIC CQI报告

    公开(公告)号:US20170041120A1

    公开(公告)日:2017-02-09

    申请号:US15106739

    申请日:2015-01-07

    摘要: A method, an apparatus, and a computer program product for wireless communication are provided. The apparatus determines a location of a reference subframe based on an A-CSI report uplink subframe, an A-CSI request downlink subframe, a reference delay, and a report delay. In an aspect, the reference delay is a first delay value before the A-CSI report uplink subframe, and the report delay is a second delay value between the A-CSI request downlink subframe and the A-CSI report uplink subframe. The apparatus determines a type of the reference subframe based on the location of the reference subframe and a subframe configuration, the type of the reference subframe being a flexible subframe or a fixed subframe. The apparatus measures at least one of a channel or interference based on the reference subframe and the type of the reference subframe. The apparatus sends, at the A-CSI report uplink subframe, an A-CSI report based on the at least one of the channel or the interference.

    摘要翻译: 提供了一种用于无线通信的方法,装置和计算机程序产品。 该装置基于A-CSI报告上行链路子帧,A-CSI请求下行链路子帧,参考延迟和报告延迟来确定参考子帧的位置。 在一方面,参考延迟是在A-CSI报告上行链路子帧之前的第一延迟值,并且报告延迟是A-CSI请求下行链路子帧和A-CSI报告上行链路子帧之间的第二延迟值。 该装置基于参考子帧的位置和子帧配置,参考子帧的类型是柔性子帧或固定子帧来确定参考子帧的类型。 该装置基于参考子帧和参考子帧的类型来测量信道或干扰中的至少一个。 该装置在A-CSI报告上行链路子帧处基于该信道或干扰中的至少一个发送A-CSI报告。

    DEVICE-TO-DEVICE FOR INTERFERENCE MITIGATION
    59.
    发明申请
    DEVICE-TO-DEVICE FOR INTERFERENCE MITIGATION 审中-公开
    用于干扰减轻的设备到设备

    公开(公告)号:US20160029396A1

    公开(公告)日:2016-01-28

    申请号:US14775335

    申请日:2014-04-09

    摘要: Certain aspects of the present disclosure relate to methods and apparatus for measuring and mitigating interference. Such interference may include base station to base station (eNB to eNB) interference and/or UE to UE interference. A base station may determine a first subframe configuration for the base station and a second subframe configuration for one or more other base stations, wherein the first and second subframe configurations have different ratios of uplink to downlink portions. The base station may measure interference caused by the one or more other base stations based on reference signals transmitted in downlink portions of the second subframe configuration.

    摘要翻译: 本公开的某些方面涉及用于测量和减轻干扰的方法和装置。 这种干扰可能包括基站到基站(eNB到eNB)的干扰和/或UE到UE的干扰。 基站可以确定用于基站的第一子帧配置和用于一个或多个其他基站的第二子帧配置,其中第一和第二子帧配置具有不同的上行链路与下行链路部分的比率。 基站可以基于在第二子帧配置的下行链路部分中发送的参考信号来测量由一个或多个其他基站引起的干扰。