DE112020002547T5 - Rausch- und signalverwaltung für rpu-array - Google Patents

Rausch- und signalverwaltung für rpu-array Download PDF

Info

Publication number
DE112020002547T5
DE112020002547T5 DE112020002547.0T DE112020002547T DE112020002547T5 DE 112020002547 T5 DE112020002547 T5 DE 112020002547T5 DE 112020002547 T DE112020002547 T DE 112020002547T DE 112020002547 T5 DE112020002547 T5 DE 112020002547T5
Authority
DE
Germany
Prior art keywords
normalizers
values
standard deviation
mean
post
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.)
Pending
Application number
DE112020002547.0T
Other languages
German (de)
English (en)
Inventor
Malte Johannes Rasch
Tayfun Gokmen
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 DE112020002547T5 publication Critical patent/DE112020002547T5/de
Pending legal-status Critical Current

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
    • G06N3/06—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
    • G06N3/065—Analogue 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)
  • Computing Systems (AREA)
  • Software Systems (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computational Linguistics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Data Mining & Analysis (AREA)
  • Artificial Intelligence (AREA)
  • Neurology (AREA)
  • Complex Calculations (AREA)
  • Magnetic Resonance Imaging Apparatus (AREA)
  • Medicines Containing Antibodies Or Antigens For Use As Internal Diagnostic Agents (AREA)
  • Radar Systems Or Details Thereof (AREA)
  • Measurement Of Velocity Or Position Using Acoustic Or Ultrasonic Waves (AREA)
DE112020002547.0T 2019-05-31 2020-04-09 Rausch- und signalverwaltung für rpu-array Pending DE112020002547T5 (de)

Applications Claiming Priority (3)

Application Number Priority Date Filing Date Title
US16/427,559 2019-05-31
US16/427,559 US11361218B2 (en) 2019-05-31 2019-05-31 Noise and signal management for RPU array
PCT/IB2020/053407 WO2020240288A1 (en) 2019-05-31 2020-04-09 Noise and signal management for rpu array

Publications (1)

Publication Number Publication Date
DE112020002547T5 true DE112020002547T5 (de) 2022-03-03

Family

ID=73550855

Family Applications (1)

Application Number Title Priority Date Filing Date
DE112020002547.0T Pending DE112020002547T5 (de) 2019-05-31 2020-04-09 Rausch- und signalverwaltung für rpu-array

Country Status (6)

Country Link
US (1) US11361218B2 (https=)
JP (1) JP7357080B2 (https=)
CN (1) CN113841164B (https=)
DE (1) DE112020002547T5 (https=)
GB (1) GB2597232B (https=)
WO (1) WO2020240288A1 (https=)

Families Citing this family (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US11482133B2 (en) * 2019-09-04 2022-10-25 Optum Services (Ireland) Limited Automatically modifying display presentations to programmatically accommodate for visual impairments
DE102019213898A1 (de) * 2019-09-11 2021-03-11 Robert Bosch Gmbh Robustes und besser trainierbares künstliches neuronales Netzwerk
US12488250B2 (en) 2020-11-02 2025-12-02 International Business Machines Corporation Weight repetition on RPU crossbar arrays
US12169534B2 (en) * 2020-12-07 2024-12-17 International Business Machines Corporation Worst case noise and bound management for RPU crossbar arrays
US12112264B2 (en) 2020-12-15 2024-10-08 International Business Machines Corporation Dynamic configuration of readout circuitry for different operations in analog resistive crossbar array
US12585940B2 (en) * 2021-09-25 2026-03-24 International Business Machines Corporation Learning static bound management parameters for analog resistive processing unit system
US20240086677A1 (en) * 2022-09-12 2024-03-14 International Business Machines Corporation Learned column-weights for rapid-estimation of properties of an entire excitation vector
CN118487605B (zh) * 2024-05-08 2025-10-14 北京犀灵视觉科技有限公司 感存算融合处理场景的模拟域信号归一化方法及装置

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9646243B1 (en) 2016-09-12 2017-05-09 International Business Machines Corporation Convolutional neural networks using resistive processing unit array
US20180293209A1 (en) 2017-04-05 2018-10-11 International Business Machines Corporation Noise and bound management for rpu array

Family Cites Families (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5500905A (en) 1991-06-12 1996-03-19 Microelectronics And Computer Technology Corporation Pattern recognition neural network with saccade-like operation
CA2081036C (en) 1991-12-19 1999-08-03 Ahmed Hashem Abdelmonem Method and apparatus for predicting transmission system errors and failures
US6078843A (en) 1997-01-24 2000-06-20 Honeywell Inc. Neural network including input normalization for use in a closed loop control system
US6054710A (en) 1997-12-18 2000-04-25 Cypress Semiconductor Corp. Method and apparatus for obtaining two- or three-dimensional information from scanning electron microscopy
EP3483795B1 (en) * 2015-01-28 2021-03-10 Google LLC Batch normalization layers
US10373054B2 (en) 2015-04-19 2019-08-06 International Business Machines Corporation Annealed dropout training of neural networks
JP6523815B2 (ja) * 2015-06-22 2019-06-05 株式会社日立製作所 プラント診断装置及びプラント診断方法
US10248907B2 (en) * 2015-10-20 2019-04-02 International Business Machines Corporation Resistive processing unit
WO2017214507A1 (en) * 2016-06-09 2017-12-14 Progress, Inc. Neural network and method of neural network training
US10755170B2 (en) * 2017-03-01 2020-08-25 International Business Machines Corporation Resistive processing unit with hysteretic updates for neural network training
US10990874B2 (en) 2017-05-22 2021-04-27 Sap Se Predicting wildfires on the basis of biophysical indicators and spatiotemporal properties using a convolutional neural network
US10192161B1 (en) 2017-12-13 2019-01-29 International Business Machines Corporation Lithium-drift based resistive processing unit for accelerating machine learning training
US11947668B2 (en) * 2018-10-12 2024-04-02 Sophos Limited Methods and apparatus for preserving information between layers within a neural network
EP3921796B1 (en) * 2019-02-06 2026-04-08 The University of British Columbia Neural network image analysis

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9646243B1 (en) 2016-09-12 2017-05-09 International Business Machines Corporation Convolutional neural networks using resistive processing unit array
US20180293209A1 (en) 2017-04-05 2018-10-11 International Business Machines Corporation Noise and bound management for rpu array

Also Published As

Publication number Publication date
US20200380348A1 (en) 2020-12-03
GB2597232B (en) 2023-02-15
JP2022534380A (ja) 2022-07-29
JP7357080B2 (ja) 2023-10-05
CN113841164B (zh) 2023-05-16
GB2597232A (en) 2022-01-19
US11361218B2 (en) 2022-06-14
WO2020240288A1 (en) 2020-12-03
GB202116922D0 (en) 2022-01-05
CN113841164A (zh) 2021-12-24

Similar Documents

Publication Publication Date Title
DE112020002547T5 (de) Rausch- und signalverwaltung für rpu-array
DE112016003245B4 (de) Resistive Verarbeitungseinheit
DE112020002186B4 (de) Dnn-training mit asymmetrischen rpu-einheiten
DE112018004223B4 (de) Trainieren künstlicher neuronaler Netze
DE112018000723T5 (de) Aktualisierungsverwaltung für eine RPU-Anordnung
DE68928385T2 (de) Neuronenrechner
DE112020005610T5 (de) Identifizieren von optimalen gewichtungen zum verbessern einervorhersagegenauigkeit bei methoden für maschinelles lernen
DE112018006189T5 (de) Robuste gradienten-gewichtskomprimierungsschemata für deep-learning-anwendungen
DE102019116305A1 (de) Pipelining zur verbesserung der inferenzgenauigkeit neuronaler netze
DE112018004992B4 (de) Übertragung synaptischer gewichte zwischen leitfähigkeitspaaren mitpolaritätsumkehr zum verringern fester einheitenasymmetrien
DE112018005227T5 (de) Merkmalsextraktion mithilfe von multi-task-lernen
DE202017007641U1 (de) Training von maschinellen Lernmodellen auf mehrere maschinelle Lernaufgaben
DE112011101370T5 (de) Neuronales Netz mit kanonischen gepulsten Neuronen für einen raumzeitlichen Assoziativspeicher
DE112018004693T5 (de) Verbessern der effizienz eines neuronalen netzes
DE112019003529T5 (de) Datendarstellung für dynamische Genauigkeit in Kernen neuronaler Netze
DE112018002950T5 (de) System und Verfahren zum Aufbauen von synaptischen Gewichten für künstliche neuronale Netze aus mit Vorzeichen versehenen analogen Leitwertpaaren unterschiedlicher Wertigkeit
DE112020003050T5 (de) Fehlerkompensation in analogen neuronalen netzen
DE112021005637T5 (de) Worst-case-rausch- und grenzmanagement bei rpu-kreuzschienenanordnungen
DE102018127802A1 (de) Hybrider klassifikator eines gepulsten neuronalen netzwerks und einer support-vektor-maschine
DE112021006672T5 (de) Dnn-training durch filtern mit versteckter matrix
DE112021001968T5 (de) Driftregularisierung zum entgegenwirken einer variation in driftkoeffizienten für analoge beschleuniger
DE112021004664T5 (de) Gattergesteuerte einheit für ein gattergesteuertes rekurrentes neuronales netzwerk
DE69517249T2 (de) Anordnung zur segmentierung eines diskreten datensatzes
DE112021004941B4 (de) Gewichtungswiederholung in rpu-kreuzschienen-arrays
DE112020005613T5 (de) Neuromorphe Einheit mit Kreuzschienen-Array-Struktur

Legal Events

Date Code Title Description
R012 Request for examination validly filed