CN108431832B - 利用外部存储器扩增神经网络 - Google Patents

利用外部存储器扩增神经网络 Download PDF

Info

Publication number
CN108431832B
CN108431832B CN201680072537.6A CN201680072537A CN108431832B CN 108431832 B CN108431832 B CN 108431832B CN 201680072537 A CN201680072537 A CN 201680072537A CN 108431832 B CN108431832 B CN 108431832B
Authority
CN
China
Prior art keywords
neural network
external memory
write
weight
read
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201680072537.6A
Other languages
English (en)
Chinese (zh)
Other versions
CN108431832A (zh
Inventor
亚历山大·本杰明·格拉韦斯
伊沃·达尼赫尔卡
蒂莫西·詹姆斯·亚历山大·哈莱
马尔科尔姆·凯文·坎贝尔·雷诺兹
格雷戈里·邓肯·韦恩
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
GDM Holding LLC
Original Assignee
DeepMind Technologies Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by DeepMind Technologies Ltd filed Critical DeepMind Technologies Ltd
Publication of CN108431832A publication Critical patent/CN108431832A/zh
Application granted granted Critical
Publication of CN108431832B publication Critical patent/CN108431832B/zh
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/082—Learning methods modifying the architecture, e.g. adding, deleting or silencing nodes or connections
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/044—Recurrent networks, e.g. Hopfield networks
    • G06N3/0442—Recurrent networks, e.g. Hopfield networks characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU]
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/0499—Feedforward networks
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/09—Supervised learning
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/092—Reinforcement learning
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/06—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons
    • G06N3/063—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons using electronic means

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Health & Medical Sciences (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • General Health & Medical Sciences (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • Computational Linguistics (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Artificial Intelligence (AREA)
  • Neurology (AREA)
  • Image Analysis (AREA)
  • Image Processing (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
CN201680072537.6A 2015-12-10 2016-12-09 利用外部存储器扩增神经网络 Active CN108431832B (zh)

Applications Claiming Priority (3)

Application Number Priority Date Filing Date Title
US201562265912P 2015-12-10 2015-12-10
US62/265,912 2015-12-10
PCT/US2016/066020 WO2017100711A1 (en) 2015-12-10 2016-12-09 Augmenting neural networks with external memory

Publications (2)

Publication Number Publication Date
CN108431832A CN108431832A (zh) 2018-08-21
CN108431832B true CN108431832B (zh) 2022-09-13

Family

ID=57708799

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201680072537.6A Active CN108431832B (zh) 2015-12-10 2016-12-09 利用外部存储器扩增神经网络

Country Status (6)

Country Link
US (3) US10832134B2 (enExample)
EP (1) EP3371747B1 (enExample)
JP (1) JP6651629B2 (enExample)
KR (1) KR102158683B1 (enExample)
CN (1) CN108431832B (enExample)
WO (1) WO2017100711A1 (enExample)

Families Citing this family (20)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10044751B2 (en) * 2015-12-28 2018-08-07 Arbor Networks, Inc. Using recurrent neural networks to defeat DNS denial of service attacks
US20180189266A1 (en) * 2017-01-03 2018-07-05 Wipro Limited Method and a system to summarize a conversation
WO2018142378A1 (en) * 2017-02-06 2018-08-09 Deepmind Technologies Limited Memory augmented generative temporal models
US11037330B2 (en) * 2017-04-08 2021-06-15 Intel Corporation Low rank matrix compression
US11243944B2 (en) * 2017-06-29 2022-02-08 Futurewei Technologies, Inc. Dynamic semantic networks for language understanding and question answering
CN107508866B (zh) * 2017-08-08 2020-10-02 重庆大学 减小移动设备端神经网络模型更新的传输消耗的方法
US11188820B2 (en) 2017-09-08 2021-11-30 International Business Machines Corporation Deep neural network performance analysis on shared memory accelerator systems
US10853725B2 (en) * 2018-05-18 2020-12-01 Deepmind Technologies Limited Neural networks with relational memory
JP6906478B2 (ja) * 2018-05-23 2021-07-21 株式会社東芝 情報処理装置、情報処理方法、およびプログラム
KR102788532B1 (ko) 2018-05-30 2025-03-31 삼성전자주식회사 뉴럴 네트워크 시스템, 이를 포함하는 어플리케이션 프로세서 및 뉴럴 네트워크 시스템의 동작방법
US11775815B2 (en) * 2018-08-10 2023-10-03 Samsung Electronics Co., Ltd. System and method for deep memory network
US20200090035A1 (en) * 2018-09-19 2020-03-19 International Business Machines Corporation Encoder-decoder memory-augmented neural network architectures
WO2020064988A1 (en) 2018-09-27 2020-04-02 Deepmind Technologies Limited Scalable and compressive neural network data storage system
CN110377342B (zh) * 2019-06-10 2022-08-30 平安科技（深圳）有限公司 基于卷积神经网络的显存处理方法、装置及存储介质
CN112149049A (zh) 2019-06-26 2020-12-29 北京百度网讯科技有限公司 用于变换矩阵的装置和方法、数据处理系统
US11651209B1 (en) * 2019-10-02 2023-05-16 Google Llc Accelerated embedding layer computations
US11461645B2 (en) * 2019-12-02 2022-10-04 International Business Machines Corporation Initialization of memory networks
US11604976B2 (en) 2020-04-29 2023-03-14 International Business Machines Corporation Crossbar arrays for computations in memory-augmented neural networks
US20220245217A1 (en) * 2021-01-29 2022-08-04 Oracle International Corporation Adaptive Selection of Source Matrix Version for Matrix Multiply Operations
US20260072786A1 (en) * 2024-09-10 2026-03-12 SK Hynix Inc. Read threshold prediction based on weight-sharing deep neural networks in non-volatile memory devices

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101253493A (zh) * 2005-08-31 2008-08-27 微软公司 在图形处理单元上训练卷积神经网络
CN104657776A (zh) * 2013-11-22 2015-05-27 华为技术有限公司 神经网络系统、基于神经网络系统的图像解析方法和装置

Family Cites Families (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5167006A (en) * 1989-12-29 1992-11-24 Ricoh Company, Ltd. Neuron unit, neural network and signal processing method
US6092018A (en) * 1996-02-05 2000-07-18 Ford Global Technologies, Inc. Trained neural network engine idle speed control system
AU5804100A (en) * 1999-05-24 2000-12-12 Ip Century Ag Neuronal network for computer-assisted knowledge management
US20050041453A1 (en) * 2003-08-22 2005-02-24 Brazis Paul W. Method and apparatus for reading and writing to solid-state memory
GB0426982D0 (en) * 2004-12-09 2005-01-12 Secr Defence Early detection of sepsis
US8239611B2 (en) * 2007-12-28 2012-08-07 Spansion Llc Relocating data in a memory device
US9514739B2 (en) * 2012-06-06 2016-12-06 Cypress Semiconductor Corporation Phoneme score accelerator
US9141906B2 (en) * 2013-03-13 2015-09-22 Google Inc. Scoring concept terms using a deep network
CN103617235B (zh) * 2013-11-26 2017-01-25 中国科学院信息工程研究所 一种基于粒子群算法的网络水军账号识别方法及系统
US10366327B2 (en) * 2014-01-31 2019-07-30 Google Llc Generating vector representations of documents
CN103824291B (zh) * 2014-02-24 2017-01-11 哈尔滨工程大学 连续量子雁群算法演化脉冲耦合神经网络系统参数的自动图像分割方法

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101253493A (zh) * 2005-08-31 2008-08-27 微软公司 在图形处理单元上训练卷积神经网络
CN104657776A (zh) * 2013-11-22 2015-05-27 华为技术有限公司 神经网络系统、基于神经网络系统的图像解析方法和装置

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
Alex Graves 等.Neural Turing Machines.《arXiv》.2014, *
MEMORY NETWORKS;Jason Weston 等;《arXiv》;20151129;正文第1-2页第2节 *
Neural Turing Machines;Alex Graves 等;《arXiv》;20141210;正文第2.3，3，3.1，3.2，3.3，3.3.1，3.3.2节，附图1-2 *
Structured Memory for Neural Turing Machines;Wei Zhang 等;《arXiv》;20151025;1-4 *

Also Published As

Publication number Publication date
US20170169332A1 (en) 2017-06-15
WO2017100711A1 (en) 2017-06-15
CN108431832A (zh) 2018-08-21
KR20180091850A (ko) 2018-08-16
US20250315676A1 (en) 2025-10-09
KR102158683B1 (ko) 2020-09-22
JP2018537788A (ja) 2018-12-20
JP6651629B2 (ja) 2020-02-19
EP3371747B1 (en) 2023-07-19
US10832134B2 (en) 2020-11-10
EP3371747A1 (en) 2018-09-12
US20210117801A1 (en) 2021-04-22
US12299575B2 (en) 2025-05-13

Similar Documents

Publication Publication Date Title
CN108431832B (zh) 利用外部存储器扩增神经网络
US11210579B2 (en) Augmenting neural networks with external memory
US11080594B2 (en) Augmenting neural networks with external memory using reinforcement learning
CN108351982B (zh) 卷积门控递归神经网络
CN107145940B (zh) 压缩的递归神经网络模型
US12260334B2 (en) Neural programming
CN109313720B (zh) 具有稀疏访问的外部存储器的增强神经网络
US12099928B2 (en) Augmented recurrent neural network with external memory
EP3238144B1 (en) Augmenting neural networks to generate additional outputs
CN109155002A (zh) 具有外部存储器的增强神经网络
CN109196527A (zh) 广度和深度机器学习模型
CN110770759A (zh) 神经网络系统

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
TR01 Transfer of patent right
TR01 Transfer of patent right

Effective date of registration: 20260120

Address after: U.S.A.

Patentee after: GDM Holdings Ltd.

Country or region after: U.S.A.

Address before: London, England

Patentee before: DEEPMIND TECHNOLOGIES Ltd.

Country or region before: United Kingdom