JP7163381B2 - ニューラル・ネットワークの効率の促進 - Google Patents

ニューラル・ネットワークの効率の促進 Download PDF

Info

Publication number
JP7163381B2
JP7163381B2 JP2020521465A JP2020521465A JP7163381B2 JP 7163381 B2 JP7163381 B2 JP 7163381B2 JP 2020521465 A JP2020521465 A JP 2020521465A JP 2020521465 A JP2020521465 A JP 2020521465A JP 7163381 B2 JP7163381 B2 JP 7163381B2
Authority
JP
Japan
Prior art keywords
activation function
component
training
computer
neural network
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
JP2020521465A
Other languages
English (en)
Japanese (ja)
Other versions
JP2021500654A5 (enExample
JP2021500654A (ja
Inventor
ワン、ヂオ
チェ、ジンウク
ゴパラクリシュナン、カイラッシュ
ヴェンカタマニ、スワガス
サカ、チャーベル
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.)
International Business Machines Corp
Original Assignee
International Business Machines Corp
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 International Business Machines Corp filed Critical International Business Machines Corp
Publication of JP2021500654A publication Critical patent/JP2021500654A/ja
Publication of JP2021500654A5 publication Critical patent/JP2021500654A5/ja
Application granted granted Critical
Publication of JP7163381B2 publication Critical patent/JP7163381B2/ja
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
    • G06N3/084—Backpropagation, e.g. using gradient descent
    • 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
    • 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
    • G06N3/0464—Convolutional networks [CNN, ConvNet]
    • 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
    • G06N3/048—Activation functions
    • 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
    • G06N3/0495—Quantised networks; Sparse networks; Compressed 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
    • 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
    • G06N3/06—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)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Computing Systems (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Molecular Biology (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Feedback Control In General (AREA)
  • Compression, Expansion, Code Conversion, And Decoders (AREA)
  • Image Analysis (AREA)
JP2020521465A 2017-10-24 2018-10-04 ニューラル・ネットワークの効率の促進 Active JP7163381B2 (ja)

Applications Claiming Priority (3)

Application Number Priority Date Filing Date Title
US15/792,733 2017-10-24
US15/792,733 US11195096B2 (en) 2017-10-24 2017-10-24 Facilitating neural network efficiency
PCT/IB2018/057712 WO2019082005A1 (en) 2017-10-24 2018-10-04 FACILITATING THE EFFECTIVENESS OF ARTIFICIAL NEURONIC NETWORK

Publications (3)

Publication Number Publication Date
JP2021500654A JP2021500654A (ja) 2021-01-07
JP2021500654A5 JP2021500654A5 (enExample) 2022-06-07
JP7163381B2 true JP7163381B2 (ja) 2022-10-31

Family

ID=66169344

Family Applications (1)

Application Number Title Priority Date Filing Date
JP2020521465A Active JP7163381B2 (ja) 2017-10-24 2018-10-04 ニューラル・ネットワークの効率の促進

Country Status (6)

Country Link
US (1) US11195096B2 (enExample)
JP (1) JP7163381B2 (enExample)
CN (1) CN111226233A (enExample)
DE (1) DE112018004693T5 (enExample)
GB (1) GB2581728A (enExample)
WO (1) WO2019082005A1 (enExample)

Families Citing this family (22)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11195096B2 (en) * 2017-10-24 2021-12-07 International Business Machines Corporation Facilitating neural network efficiency
US20210166106A1 (en) * 2017-12-12 2021-06-03 The Regents Of The University Of California Residual binary neural network
US12536418B2 (en) * 2018-04-27 2026-01-27 Carnegie Mellon University Perturbative neural network
JP7287388B2 (ja) * 2018-05-14 2023-06-06 ソニーグループ株式会社 情報処理装置および情報処理方法
US11763133B2 (en) * 2018-08-31 2023-09-19 Servicenow Canada Inc. Data point suitability determination from edge device neural networks
KR102621118B1 (ko) * 2018-11-01 2024-01-04 삼성전자주식회사 영상 적응적 양자화 테이블을 이용한 영상의 부호화 장치 및 방법
US20200302276A1 (en) * 2019-03-20 2020-09-24 Gyrfalcon Technology Inc. Artificial intelligence semiconductor chip having weights of variable compression ratio
US20220284300A1 (en) * 2019-09-19 2022-09-08 Intel Corporation Techniques to tune scale parameter for activations in binary neural networks
JP7419035B2 (ja) * 2019-11-22 2024-01-22 キヤノン株式会社 学習モデル管理システム、学習モデル管理方法、およびプログラム
US20210174214A1 (en) * 2019-12-10 2021-06-10 The Mathworks, Inc. Systems and methods for quantizing a neural network
US11935271B2 (en) * 2020-01-10 2024-03-19 Tencent America LLC Neural network model compression with selective structured weight unification
US11823054B2 (en) 2020-02-20 2023-11-21 International Business Machines Corporation Learned step size quantization
WO2021211099A1 (en) * 2020-04-14 2021-10-21 Google Llc Efficient binary representations from neural networks
CN113762452B (zh) * 2020-06-04 2024-01-02 合肥君正科技有限公司 一种量化prelu激活函数的方法
CN113778655B (zh) * 2020-06-09 2025-01-24 北京灵汐科技有限公司 一种网络精度的量化方法及系统
US20220079491A1 (en) * 2020-09-14 2022-03-17 Biosense Webster (Israel) Ltd. Local activation time analysis system
US12400112B2 (en) * 2020-12-08 2025-08-26 International Business Machines Corporation Efficient method for VLSI implementation of useful neural network activation functions
CN114692817A (zh) * 2020-12-31 2022-07-01 合肥君正科技有限公司 一种动态调整量化feature clip值的方法
CN112749803B (zh) * 2021-03-05 2023-05-30 成都启英泰伦科技有限公司 一种神经网络的激活函数计算量化方法
WO2022216109A1 (en) * 2021-04-09 2022-10-13 Samsung Electronics Co., Ltd. Method and electronic device for quantizing dnn model
JP7700577B2 (ja) 2021-08-25 2025-07-01 富士通株式会社 閾値決定プログラム、閾値決定方法および閾値決定装置
KR102650510B1 (ko) * 2022-10-28 2024-03-22 한국전자기술연구원 영상의 노이즈 제거 방법 및 장치

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160328646A1 (en) 2015-05-08 2016-11-10 Qualcomm Incorporated Fixed point neural network based on floating point neural network quantization
CN107122825A (zh) 2017-03-09 2017-09-01 华南理工大学 一种神经网络模型的激活函数生成方法

Family Cites Families (33)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
KR950012359B1 (ko) * 1992-08-28 1995-10-17 엘지전자주식회사 신경회로망 구조와 학습방법
JPH1063633A (ja) * 1996-08-26 1998-03-06 Denso Corp ニューラルネットワークの演算装置及び車両用空調装置
US7149262B1 (en) * 2000-07-06 2006-12-12 The Trustees Of Columbia University In The City Of New York Method and apparatus for enhancing data resolution
HUP0301368A3 (en) * 2003-05-20 2005-09-28 Amt Advanced Multimedia Techno Method and equipment for compressing motion picture data
US20150120627A1 (en) 2013-10-29 2015-04-30 Qualcomm Incorporated Causal saliency time inference
US20150269480A1 (en) 2014-03-21 2015-09-24 Qualcomm Incorporated Implementing a neural-network processor
US10417525B2 (en) * 2014-09-22 2019-09-17 Samsung Electronics Co., Ltd. Object recognition with reduced neural network weight precision
US20170032247A1 (en) * 2015-07-31 2017-02-02 Qualcomm Incorporated Media classification
US11029949B2 (en) * 2015-10-08 2021-06-08 Shanghai Zhaoxin Semiconductor Co., Ltd. Neural network unit
KR102565273B1 (ko) 2016-01-26 2023-08-09 삼성전자주식회사 뉴럴 네트워크에 기초한 인식 장치 및 뉴럴 네트워크의 학습 방법
US10831444B2 (en) * 2016-04-04 2020-11-10 Technion Research & Development Foundation Limited Quantized neural network training and inference
US10621486B2 (en) * 2016-08-12 2020-04-14 Beijing Deephi Intelligent Technology Co., Ltd. Method for optimizing an artificial neural network (ANN)
US11003985B2 (en) * 2016-11-07 2021-05-11 Electronics And Telecommunications Research Institute Convolutional neural network system and operation method thereof
US10373049B2 (en) * 2016-12-20 2019-08-06 Google Llc Generating an output for a neural network output layer
KR102457463B1 (ko) * 2017-01-16 2022-10-21 한국전자통신연구원 희소 파라미터를 사용하는 압축 신경망 시스템 및 그것의 설계 방법
US10127495B1 (en) * 2017-04-14 2018-11-13 Rohan Bopardikar Reducing the size of a neural network through reduction of the weight matrices
CN107229942B (zh) 2017-04-16 2021-03-30 北京工业大学 一种基于多个分类器的卷积神经网络分类方法
US20180336469A1 (en) * 2017-05-18 2018-11-22 Qualcomm Incorporated Sigma-delta position derivative networks
EP3657398B1 (en) * 2017-05-23 2025-10-08 Shanghai Cambricon Information Technology Co., Ltd Processing method and accelerating device
KR102526650B1 (ko) * 2017-05-25 2023-04-27 삼성전자주식회사 뉴럴 네트워크에서 데이터를 양자화하는 방법 및 장치
US10878273B2 (en) * 2017-07-06 2020-12-29 Texas Instruments Incorporated Dynamic quantization for deep neural network inference system and method
US10728553B2 (en) * 2017-07-11 2020-07-28 Sony Corporation Visual quality preserving quantization parameter prediction with deep neural network
KR102601604B1 (ko) * 2017-08-04 2023-11-13 삼성전자주식회사 뉴럴 네트워크의 파라미터들을 양자화하는 방법 및 장치
US10839286B2 (en) * 2017-09-14 2020-11-17 Xilinx, Inc. System and method for implementing neural networks in integrated circuits
KR102728799B1 (ko) * 2017-09-25 2024-11-11 삼성전자주식회사 인공 신경망의 양자화 방법 및 장치
US11195096B2 (en) * 2017-10-24 2021-12-07 International Business Machines Corporation Facilitating neural network efficiency
US11132605B2 (en) * 2017-11-20 2021-09-28 International Business Machines Corporation Cardinal sine as an activation function for universal classifier training data
US11295208B2 (en) * 2017-12-04 2022-04-05 International Business Machines Corporation Robust gradient weight compression schemes for deep learning applications
US11551077B2 (en) * 2018-06-13 2023-01-10 International Business Machines Corporation Statistics-aware weight quantization
US20200226459A1 (en) * 2019-01-11 2020-07-16 International Business Machines Corporation Adversarial input identification using reduced precision deep neural networks
US11551054B2 (en) * 2019-08-27 2023-01-10 International Business Machines Corporation System-aware selective quantization for performance optimized distributed deep learning
US12175359B2 (en) * 2019-09-03 2024-12-24 International Business Machines Corporation Machine learning hardware having reduced precision parameter components for efficient parameter update
US20210125063A1 (en) * 2019-10-23 2021-04-29 Electronics And Telecommunications Research Institute Apparatus and method for generating binary neural network

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20160328646A1 (en) 2015-05-08 2016-11-10 Qualcomm Incorporated Fixed point neural network based on floating point neural network quantization
CN107122825A (zh) 2017-03-09 2017-09-01 华南理工大学 一种神经网络模型的激活函数生成方法

Also Published As

Publication number Publication date
CN111226233A (zh) 2020-06-02
US11195096B2 (en) 2021-12-07
DE112018004693T5 (de) 2020-06-18
WO2019082005A1 (en) 2019-05-02
US20190122116A1 (en) 2019-04-25
GB202006969D0 (en) 2020-06-24
JP2021500654A (ja) 2021-01-07
GB2581728A (en) 2020-08-26

Similar Documents

Publication Publication Date Title
JP7163381B2 (ja) ニューラル・ネットワークの効率の促進
US11334671B2 (en) Adding adversarial robustness to trained machine learning models
KR102877638B1 (ko) 뉴럴 네트워크 양자화를 위한 방법 및 장치
TWI791610B (zh) 對人工神經網路及浮點神經網路進行量化的方法及裝置
CN110598838A (zh) 统计感知权重量化
JP7316725B2 (ja) エンコーダ－デコーダ・メモリ拡張ニューラル・ネットワーク・アーキテクチャ
US11468313B1 (en) Systems and methods for quantizing neural networks via periodic regularization functions
WO2020113355A1 (en) A content adaptive attention model for neural network-based image and video encoders
Hammad Deep learning activation functions: Fixed-shape, parametric, adaptive, stochastic, miscellaneous, non-standard, ensemble
JP7083438B2 (ja) ニューラル・ネットワークにおけるシナプス活性のための自動調整閾値
US20220284298A1 (en) Method and apparatus for pruning neural networks
US12585933B2 (en) Transfer learning with augmented neural networks
JP2020517002A (ja) ニューラル・ネットワークの更新管理のためのコンピュータ実装方法、コンピュータ・プログラム、およびコンピュータ処理システム
JP7149503B2 (ja) 階層型ニューラル・ネットワークにおけるフィードバックのための計算方法
CN114257313A (zh) 基于深度学习的光纤光栅传感系统降噪方法及其相关设备
US11341598B2 (en) Interpretation maps with guaranteed robustness
Li et al. SEMQ: Efficient non-uniform quantization with sensitivity-based error minimization for large language models
KR20200048002A (ko) 비대칭 tanh 활성 함수를 이용한 예측 성능의 개선
US20190258928A1 (en) Artificial neural network
CN119806793A (zh) 模拟内存计算系统的自适应量化方法、自适应量化系统和非暂时性机器可读介质
WO2019121142A1 (en) Approaching homeostasis in a binary neural network
CN117952174A (zh) 神经网络剪枝方法及相关装置
KR20240030698A (ko) 라그랑지함수 기반 에러역전파를 이용한 심층신경망 가중치 양자화 방법
US12265911B2 (en) Neural network layers with a controlled degree of spatial invariance
US20250139422A1 (en) Method and device with neural network bit-width quantization training

Legal Events

Date Code Title Description
A521 Request for written amendment filed

Free format text: JAPANESE INTERMEDIATE CODE: A523

Effective date: 20210309

A621 Written request for application examination

Free format text: JAPANESE INTERMEDIATE CODE: A621

Effective date: 20210323

A977 Report on retrieval

Free format text: JAPANESE INTERMEDIATE CODE: A971007

Effective date: 20220218

A131 Notification of reasons for refusal

Free format text: JAPANESE INTERMEDIATE CODE: A131

Effective date: 20220301

RD04 Notification of resignation of power of attorney

Free format text: JAPANESE INTERMEDIATE CODE: A7424

Effective date: 20220502

A524 Written submission of copy of amendment under article 19 pct

Free format text: JAPANESE INTERMEDIATE CODE: A524

Effective date: 20220527

TRDD Decision of grant or rejection written
A01 Written decision to grant a patent or to grant a registration (utility model)

Free format text: JAPANESE INTERMEDIATE CODE: A01

Effective date: 20221011

A61 First payment of annual fees (during grant procedure)

Free format text: JAPANESE INTERMEDIATE CODE: A61

Effective date: 20221019

R150 Certificate of patent or registration of utility model

Ref document number: 7163381

Country of ref document: JP

Free format text: JAPANESE INTERMEDIATE CODE: R150