公开(公告)号：US12093573B2

公开(公告)日：2024-09-17

申请号：US17163179

申请日：2021-01-29

Applicant: Intel Corporation

Inventor： Alaa Alameldeen ,  Amir Ali Radjai ,  Jason Van Dyken ,  Aditya Nagaraja ,  Joshua Underdown ,  James Greensky ,  Wei Wu

IPC: G06F3/06 ,  G06F11/10 ,  H03M7/30

CPC classification number: G06F3/0661 ,  G06F3/0608 ,  G06F3/0673 ,  G06F11/1076 ,  H03M7/30

Abstract: An embodiment of an electronic apparatus may include a substrate, and logic coupled to the substrate, the logic to determine a base value to compress a block of data, wherein the block of data consists of a first number of data words, replace original values from a second number of data words from the block of data with respective delta values from the base value to provide compressed data, wherein the second number of data words is at least two less than the first number of data words, and store metadata associated with the block of data together with the compressed data in the block of data. Other embodiments are disclosed and claimed.

公开(公告)号：US12079722B2

公开(公告)日：2024-09-03

申请号：US18162871

申请日：2023-02-01

Applicant: TuSimple, Inc. ,  Beijing Tusen Zhitu Technology Co., Ltd.

Inventor： Yuwei Hu ,  Jiangming Jin ,  Lei Su ,  Dinghua Li

IPC: G06F12/02 ,  G06F17/15 ,  G06F17/16 ,  G06N3/045 ,  G06N3/063 ,  G06N3/08 ,  G06N20/10 ,  H03M7/30

CPC classification number: G06N3/08 ,  G06F12/0207 ,  G06F17/153 ,  G06F17/16 ,  G06N3/045 ,  G06N3/063 ,  G06N20/10 ,  H03M7/30

Abstract: The embodiments of this application provide a method and device for optimizing neural network. The method includes: binarizing and bit-packing input data of a convolution layer along a channel direction, and obtaining compressed input data; binarizing and bit-packing respectively each convolution kernel of the convolution layer along the channel direction, and obtaining each corresponding compressed convolution kernel; dividing the compressed input data sequentially in a convolutional computation order into blocks of the compressed input data with the same size of each compressed convolution kernel, wherein the data input to one time convolutional computation form a data block; and, taking a convolutional computation on each block of the compressed input data and each compressed convolution kernel sequentially, obtaining each convolutional result data, and obtaining multiple output data of the convolution layer according to each convolutional result data.

公开(公告)号：US12039985B2

公开(公告)日：2024-07-16

申请号：US18301191

申请日：2023-04-14

Applicant: Fraunhofer-Gesellschaft zur Foerderung der angewandten Forschung e.V.

Inventor： Markus Multrus ,  Bernhard Grill ,  Guillaume Fuchs ,  Stefan Geyersberger ,  Nikolaus Rettelbach ,  Virgilio Bacigalupo

IPC: G10L19/022 ,  G10L19/02 ,  G10L19/025 ,  H03M7/30

CPC classification number: G10L19/02 ,  G10L19/022 ,  H03M7/30

Abstract: An audio encoder for encoding segments of coefficients, the segments of coefficients representing different time or frequency resolutions of a sampled audio signal, the audio encoder including a processor for deriving a coding context for a currently encoded coefficient of a current segment based on a previously encoded coefficient of a previous segment, the previously encoded coefficient representing a different time or frequency resolution than the currently encoded coefficient. The audio encoder further includes an entropy encoder for entropy encoding the current coefficient based on the coding context to obtain an encoded audio stream.

公开(公告)号：US12034461B2

公开(公告)日：2024-07-09

申请号：US17795333

申请日：2020-03-06

Applicant: Telefonaktiebolaget LM Ericsson (publ)

Inventor： Henrik Rydén ,  Roy Timo ,  Maksym Girnyk ,  Hamed Farhadi

IPC: H03M7/30 ,  G01S5/00 ,  H04L25/02 ,  H04W64/00

CPC classification number: H03M7/30 ,  G01S5/0036 ,  H04L25/0212 ,  H04L25/0254 ,  H03M7/3059 ,  H03M7/70 ,  H04W64/006

Abstract: In a first method, a wireless device estimates a delay profile of a channel impulse response, CIR, for a channel between a network node and the wireless device, compresses the delay profile using a compression function, and transmits the compressed delay profile. The compression function includes a first function and a quantizer. The first function is configured to receive input data and reduce a dimension of the input data. In a second method, a network node receives a compressed delay profile of CIR for a channel between a network node and a wireless device, decompresses the compressed delay profile using a decompression function, and estimates a position of the wireless device based on at least the decompressed delay profile. The decompression function includes a first function which is configured to receive input data and provide output data in a higher dimensional space than the input data.

公开(公告)号：US20240220432A1

公开(公告)日：2024-07-04

申请号：US18601409

申请日：2024-03-11

Applicant: HUAWEI TECHNOLOGIES CO., LTD.

Inventor： Baoqing LIU ,  Hu LIU ,  Qinglong CHEN

IPC: G06F13/28 ,  G06F17/16 ,  H03M7/30

CPC classification number: G06F13/28 ,  G06F17/16 ,  H03M7/30

Abstract: The present disclosure provides example operation accelerators and compression methods. One example operation accelerator includes a storage configured to store first input data, weight data, and a control instruction, and an operation circuit connected to the storage and configured to perform matrix multiplication on the first input data and the weight data, to obtain a computation result. The operation accelerator further includes a compression module configured to compress the computation result to obtain compressed data, as well as a controller connected to the storage and configured to obtain the control instruction from the storage, and when the control instruction includes instructions to compress the computation result, control the compression module to compress the computation result to obtain the compressed data. The operation accelerator further includes a direct memory access controller connected to the compression module and configured to store the compressed data in a memory outside the operation accelerator.

公开(公告)号：US20240195433A1

公开(公告)日：2024-06-13

申请号：US18553415

申请日：2022-04-04

Applicant: Nokia Technologies Oy

Inventor： Emre Baris AKSU ,  Miska Matias HANNUKSELA ,  Hamed REZAZADEGAN TAVAKOLI ,  Francesco CRICRÌ

IPC: H03M7/30

CPC classification number: H03M7/30

Abstract: The embodiments relate to a method for encoding two or more tensors. The method comprises processing the two or more tensors having respective dimensions so that the dimensions of said two or more sensors have the same number (510); identifying which axis of each individual tensor is swappable to result in concatenable tensors around an axis of concatenation (520): reshaping the tensors so that the dimensions are modified based on the swapped axis (530): concatenating the tensors around the axis of concatenation to result in concatenated tensor (540): compressing the concatenated tensor (550); generating syntax structures for carrying concatenation and axis swapping information (560): and generating a bitstream by combining the syntax structures and the compressed concatenated tensor (570). The embodiments also relate to a method for decoding, and to apparatuses for implementing the methods.

公开(公告)号：US20240185291A1

公开(公告)日：2024-06-06

申请号：US18463143

申请日：2023-09-07

Applicant: Groupon, Inc.

Inventor： Ricardo A. ZILLERUELO-RAMOS ,  Hernan Enrique ARROYO GARCIA ,  Joe FRISBIE ,  Gaston L'HUILLIER ,  Francisco Jose LARRAIN

IPC: G06Q30/0242 ,  G06F16/174 ,  G06F16/22 ,  G06F16/951 ,  G06F16/955 ,  H03M7/30

CPC classification number: G06Q30/0246 ,  G06F16/1744 ,  G06F16/2282 ,  G06F16/951 ,  G06F16/955 ,  H03M7/30 ,  H03M7/70

Abstract: In general, embodiments of the present invention provide systems, methods and computer readable media for data record compression using graph-based techniques.

公开(公告)号：US20240143449A1

公开(公告)日：2024-05-02

申请号：US18405231

申请日：2024-01-05

Applicant: Huawei Technologies Co., Ltd.

Inventor： Keyun Chen

IPC: G06F11/14 ,  G06F16/174 ,  H03M7/30

CPC classification number: G06F11/1453 ,  G06F16/1748 ,  H03M7/30

Abstract: A data processing method includes a computing device that obtains first data, and determines a deduplication rate of the first data and a working status of the computing device. When the working status of the computing device is a busy state, and when the deduplication rate of the first data is greater than or equal to a first preset value, deduplication on the first data is performed. When the deduplication rate of the first data is lower than the first preset value, deduplication is not performed on the first data, and the first data may be directly stored.

公开(公告)号：US11960421B2

公开(公告)日：2024-04-16

申请号：US17216476

申请日：2021-03-29

Applicant: HUAWEI TECHNOLOGIES CO., LTD.

Inventor： Baoqing Liu ,  Hu Liu ,  Qinglong Chen

IPC: G06F13/28 ,  G06F17/16 ,  H03M7/30

CPC classification number: G06F13/28 ,  G06F17/16 ,  H03M7/30

Abstract: The present disclosure discloses example operation accelerators and compression methods. One example operation accelerator performs operations, including storing, in a first buffer, first input data. In a second buffer, weight data can be stored. A computation result is obtained by performing matrix multiplication on the first input data and the weight data by an operation circuit connected to the input buffer and the weight buffer. The computation result is compressed by a compression module to obtain compressed data. The compressed data can be stored into a memory outside the operation accelerator by a direct memory access controller (DMAC) connected to the compression module.

公开(公告)号：US11942101B2

公开(公告)日：2024-03-26

申请号：US18301194

申请日：2023-04-14

Applicant: Fraunhofer-Gesellschaft zur Foerderung der angewandten Forschung e.V.

Inventor： Markus Multrus ,  Bernhard Grill ,  Guillaume Fuchs ,  Stefan Geyersberger ,  Nikolaus Rettelbach ,  Virgilio Bacigalupo

IPC: G10L19/02 ,  G10L19/022 ,  H03M7/30

CPC classification number: G10L19/02 ,  G10L19/022 ,  H03M7/30

Abstract: An audio encoder for encoding segments of coefficients, the segments of coefficients representing different time or frequency resolutions of a sampled audio signal, the audio encoder including a processor for deriving a coding context for a currently encoded coefficient of a current segment based on a previously encoded coefficient of a previous segment, the previously encoded coefficient representing a different time or frequency resolution than the currently encoded coefficient. The audio encoder further includes an entropy encoder for entropy encoding the current coefficient based on the coding context to obtain an encoded audio stream.