SlideShare a Scribd company logo
1 of 30
Javier Fernández1,3
, Irune Agirre3
, Jon Perez-Cerrolaza3
,
Francisco J. Cazorla1
, Jaume Abella1,2
CONTENTS
01
02
03
CONTENTS
CONTEXTUALIZATION
PROPOSED SOLUTION
EVALUATION
04 CONCLUSIONS
3
Main Concepts
Artificial Intelligence
It has made enormous
progress, reaching near-
human accuracy in several
safety-related tasks.
Functional safety standards
IEC 61508 IEC 61513
EN 5012X ISO 26262
Example in the Automotive domain
4
Main Concepts Baseline concept
Detection of faults at runtime in
the Matrix-matrix Multiplication
Catalog of diagnostics
techniques
CUTLASS: High-performance matrix-
matrix multiplication Library
“On the Safe Deployment of Matrix Multiplication in Massively Parallel Safety-Related Systems”
Object detector application based
on CNNs (Tiny YOLO-v3)
Matrix-matrix
Multiplication (MMM)
It is the backbone of the Convolutional
Neural Networks in terms of execution
time:
• Sequential implementation: 98,5 %
• Vectorized implementation: 87%
• CUDA based implementation: 67 %
𝐴11 𝐴12
𝐴21 𝐴22
𝐴31 𝐴32
𝑋
𝐵11 𝐵12 𝐵13
𝐵21 𝐵22 𝐵23
=
𝐶11 𝐶12 𝐶13
𝐶21 𝐶22 𝐶23
𝐶31 𝐶32 𝐶33
CONTENTS
01
02
03
04
CONTENTS
CONTEXTUALIZATION
PROPOSED SOLUTION
EVALUATION
CONCLUSIONS
6
P R O P O S E D S O LU T I O N
(1)
(1)
(1)
1 Berkeley DeepDrive dataset (https://www.bdd100k.com/)
7
Stage 1
P R O P O S E D S O LU T I O N
1 Berkeley DeepDrive dataset (https://www.bdd100k.com/)
(1)
(1)
(1)
Stage 1
P R O P O S E D S O LU T I O N
8
(1)
(1)
(1)
1 Berkeley DeepDrive dataset (https://www.bdd100k.com/)
9
Stage 1
P R O P O S E D S O LU T I O N
(1)
(1)
(1)
1 Berkeley DeepDrive dataset (https://www.bdd100k.com/)
10
Stage 2
Stage 1
P R O P O S E D S O LU T I O N
(1)
(1)
(1)
1 Berkeley DeepDrive dataset (https://www.bdd100k.com/)
11
Stage 2
Stage 1
11
P R O P O S E D S O LU T I O N
(1)
(1)
(1)
1 Berkeley DeepDrive dataset (https://www.bdd100k.com/)
Stage 2
Stage 1
P R O P O S E D S O LU T I O N
1 Berkeley DeepDrive dataset (https://www.bdd100k.com/)
(1)
(1)
(1)
12
Stage 2
Stage 1
P R O P O S E D S O LU T I O N
(1)
(1)
(1)
1 Berkeley DeepDrive dataset (https://www.bdd100k.com/)
13
Stage 2
Stage 1 Stage 2
Stage 1
P R O P O S E D S O LU T I O N
(1)
(1)
(1)
1 Berkeley DeepDrive dataset (https://www.bdd100k.com/)
14
Stage 2
Stage 1
P R O P O S E D S O LU T I O N
(1)
(1)
(1)
1 Berkeley DeepDrive dataset (https://www.bdd100k.com/)
Stage 2
Stage 1
P R O P O S E D S O LU T I O N
(1)
(1)
(1)
1 Berkeley DeepDrive dataset (https://www.bdd100k.com/)
16
Stage 2
Stage 1
P R O P O S E D S O LU T I O N
17
DC computation per fault source:
B. Fault injected at the global memory level:
𝐷𝑒𝑡𝐴= (𝐵1𝑑𝑒𝑡𝐴
†
+ 𝐵3𝑑𝑒𝑡𝐴
†
) x 𝑁_𝐵𝑅𝑇1+𝐵2𝑑𝑒𝑡𝐴
∗
+𝐵4𝑑𝑒𝑡𝐴
∗
𝐷𝑒𝑡𝐵=(𝐵1𝑑𝑒𝑡𝐵
⊗
+ 𝐵2𝑑𝑒𝑡𝐵
⊗
) x 𝑁_𝐵𝐶𝑇1+𝐵3𝑑𝑒𝑡𝐵
△
+𝐵4𝑑𝑒𝑡𝐵
△
𝐷𝐶 =
𝐷𝑒𝑡𝐴 + 𝐷𝑒𝑡𝐵
𝑀 + 𝑁 𝑥 𝐾 𝑥 𝑑𝑎𝑡𝑎_𝑠𝑖𝑧𝑒
A. Faults injected at the arithmetic level or at the register level:
𝐷𝐶 =
𝑖=1
4
(𝑁𝑏𝑙𝑜𝑐𝑘𝑠𝐵𝑖
× 𝑁det _𝐵𝑖
)
𝑁𝑓𝑖
Stage 3
Stage 2
Stage 1
P R O P O S E D S O LU T I O N
1 Berkeley DeepDrive dataset (https://www.bdd100k.com/)
(1)
(1)
(1)
18
CONTENTS
01
02
03
04
CONTENTS
CONTEXTUALIZATION
PROPOSED SOLUTION
EVALUATION
CONCLUSIONS
Set-up
Matrix Multiplication
𝐴11 𝐴12
𝐴21 𝐴22
𝐴31 𝐴32
𝑋
𝐵11 𝐵12 𝐵13
𝐵21 𝐵22 𝐵23
=
𝐶11 𝐶12 𝐶13
𝐶21 𝐶22 𝐶23
𝐶31 𝐶32 𝐶33
𝑀𝑥𝐾 𝑀𝑥𝑁
𝐾𝑥𝑁
Implementation
E VA LUAT I O N
Set-up Stage 1
s
Sensibility to misclassification
E VA LUAT I O N
21
Stage 2
Set-up Stage 1
Performance impact (without compiler optimization) Performance impact (maximum compiler optimization)
E VA LUAT I O N
Performance impact:
L1 (Minimum): From 1,01 to 1,37
L3 (Maximum): From 1,002 to 1,18
22
Performance impact:
L1 (Minimum): From 1,02 to 82,5
L7 (Maximum): From 1,04 to 171,5
Stage 2
DC of each layer of Tiny Yolo-v3
E VA LUAT I O N
23
Set-up Stage 1
Stage 3
Selective protection
Remarks
Note that, while such performance impact is high, it could be reduced if diagnostics
are just executed once periodically. For example:
For the highest diagnostic coverage PI = 3,8x the CNN execution time
Process safety time = 100x a single classification task
----------------------------------------------------------------------------------------------------
PI is lower than 5 %
E VA LUAT I O N
24
Stage 2
Set-up Stage 1
CONTENTS
01
02
03
04
CONTENTS
CONCLUSIONS
CONTEXTUALIZATION
PROPOSED SOLUTION
EVALUATION
C O N C LU S I O N S
Conclusions
Conclusions
We propose a methodology to selectively protect CNNs deployed on GPUs decomposed
into three stages and demonstrate its applicability on a tiny version of an object detector,
tiny YOLO-v3. Additionally, we remark:
• For this CNN, we observe a higher tendency to misclassify (from 83,4 to 99,6%) in the
initial layers (L1-L8). However, the final layers also present lower but still high
misclassification rates (from 55,2 to 74,34%).
• For the given example, we observe that the lowest performance impact to achieve high,
medium, and low DC ranges is 3,8, 3,33, and 2,61, respectively.
26
IKERLAN
P.º José María Arizmendiarrieta, 2 - 20500 Arrasate-Mondragón
T. +34 943712400 F. +34 943796944
THANK YOU
IKERLAN
P.º José María Arizmendiarrieta, 2 - 20500 Arrasate-Mondragón
T. +34 943712400 F. +34 943796944
NAME: JAVIER FERNÁNDEZ MUÑOZ
EMAIL: JAVIER.FERNANDEZ@IKERLAN.ES
Acknowledgements:
• Ikerlan authors have received funding from the Elkartek grant project KK-
2021/00123 of the Basque government.
• BSC authors have been partially supported by the Spanish Ministry of Science
and Innovation under grant PID2019-
107255GBC21/AEI/10.13039/501100011033
Classification is correct if:
1. The central point of the box is less than 50 pixels away
2. Width and height of the boxes vary by less than 25 pixels
3. Accuracy differs by less than 15%.
Safe architectural patterns proposed:

More Related Content

Similar to ICSRS_R038.pptx

Compressed learning for time series classification
Compressed learning for time series classificationCompressed learning for time series classification
Compressed learning for time series classification學翰 施
 
DEEP LEARNING BASED BRAIN STROKE DETECTION
DEEP LEARNING BASED BRAIN STROKE DETECTIONDEEP LEARNING BASED BRAIN STROKE DETECTION
DEEP LEARNING BASED BRAIN STROKE DETECTIONIRJET Journal
 
Artificial Neural Networks (ANNS) For Prediction of California Bearing Ratio ...
Artificial Neural Networks (ANNS) For Prediction of California Bearing Ratio ...Artificial Neural Networks (ANNS) For Prediction of California Bearing Ratio ...
Artificial Neural Networks (ANNS) For Prediction of California Bearing Ratio ...IJMER
 
IRJET- A Survey on Medical Image Interpretation for Predicting Pneumonia
IRJET- A Survey on Medical Image Interpretation for Predicting PneumoniaIRJET- A Survey on Medical Image Interpretation for Predicting Pneumonia
IRJET- A Survey on Medical Image Interpretation for Predicting PneumoniaIRJET Journal
 
Making of a PD Data Acqusition System
Making of a PD Data Acqusition SystemMaking of a PD Data Acqusition System
Making of a PD Data Acqusition SystemVishal Mathur
 
IRJET- Jeevn-Net: Brain Tumor Segmentation using Cascaded U-Net & Overall...
IRJET-  	  Jeevn-Net: Brain Tumor Segmentation using Cascaded U-Net & Overall...IRJET-  	  Jeevn-Net: Brain Tumor Segmentation using Cascaded U-Net & Overall...
IRJET- Jeevn-Net: Brain Tumor Segmentation using Cascaded U-Net & Overall...IRJET Journal
 
Parallel knn on gpu architecture using opencl
Parallel knn on gpu architecture using openclParallel knn on gpu architecture using opencl
Parallel knn on gpu architecture using opencleSAT Journals
 
Parallel k nn on gpu architecture using opencl
Parallel k nn on gpu architecture using openclParallel k nn on gpu architecture using opencl
Parallel k nn on gpu architecture using opencleSAT Publishing House
 
Consequence driven decomposition of large-scale power system security analysis
Consequence driven decomposition of large-scale power system security analysisConsequence driven decomposition of large-scale power system security analysis
Consequence driven decomposition of large-scale power system security analysisUniversité de Liège (ULg)
 
Threshold adaptation and XOR accumulation algorithm for objects detection
Threshold adaptation and XOR accumulation algorithm for  objects detectionThreshold adaptation and XOR accumulation algorithm for  objects detection
Threshold adaptation and XOR accumulation algorithm for objects detectionIJECEIAES
 
"An adaptive modular approach to the mining of sensor network ...
"An adaptive modular approach to the mining of sensor network ..."An adaptive modular approach to the mining of sensor network ...
"An adaptive modular approach to the mining of sensor network ...butest
 
An efficient recovery mechanism
An efficient recovery mechanismAn efficient recovery mechanism
An efficient recovery mechanismijcsa
 
Design of Kalman filter for Airborne Applications
Design of Kalman filter for Airborne ApplicationsDesign of Kalman filter for Airborne Applications
Design of Kalman filter for Airborne ApplicationsIJERA Editor
 
캡슐 네트워크를 이용한 엔드투엔드 음성 단어 인식, 배재성(KAIST 석사과정)
캡슐 네트워크를 이용한 엔드투엔드 음성 단어 인식, 배재성(KAIST 석사과정)캡슐 네트워크를 이용한 엔드투엔드 음성 단어 인식, 배재성(KAIST 석사과정)
캡슐 네트워크를 이용한 엔드투엔드 음성 단어 인식, 배재성(KAIST 석사과정)NAVER Engineering
 
Field-programmable gate array design of image encryption and decryption usin...
Field-programmable gate array design of image encryption and  decryption usin...Field-programmable gate array design of image encryption and  decryption usin...
Field-programmable gate array design of image encryption and decryption usin...IJECEIAES
 
Deep learning for_ecg_classification
Deep learning for_ecg_classificationDeep learning for_ecg_classification
Deep learning for_ecg_classificationtriwiyantotriwiyanto
 
1st review android malware.pptx
1st review  android malware.pptx1st review  android malware.pptx
1st review android malware.pptxNambiraju
 

Similar to ICSRS_R038.pptx (20)

Compressed learning for time series classification
Compressed learning for time series classificationCompressed learning for time series classification
Compressed learning for time series classification
 
tracking.ppt
tracking.ppttracking.ppt
tracking.ppt
 
Report
ReportReport
Report
 
DEEP LEARNING BASED BRAIN STROKE DETECTION
DEEP LEARNING BASED BRAIN STROKE DETECTIONDEEP LEARNING BASED BRAIN STROKE DETECTION
DEEP LEARNING BASED BRAIN STROKE DETECTION
 
Artificial Neural Networks (ANNS) For Prediction of California Bearing Ratio ...
Artificial Neural Networks (ANNS) For Prediction of California Bearing Ratio ...Artificial Neural Networks (ANNS) For Prediction of California Bearing Ratio ...
Artificial Neural Networks (ANNS) For Prediction of California Bearing Ratio ...
 
ODVSML_Presentation
ODVSML_PresentationODVSML_Presentation
ODVSML_Presentation
 
IRJET- A Survey on Medical Image Interpretation for Predicting Pneumonia
IRJET- A Survey on Medical Image Interpretation for Predicting PneumoniaIRJET- A Survey on Medical Image Interpretation for Predicting Pneumonia
IRJET- A Survey on Medical Image Interpretation for Predicting Pneumonia
 
Making of a PD Data Acqusition System
Making of a PD Data Acqusition SystemMaking of a PD Data Acqusition System
Making of a PD Data Acqusition System
 
IRJET- Jeevn-Net: Brain Tumor Segmentation using Cascaded U-Net & Overall...
IRJET-  	  Jeevn-Net: Brain Tumor Segmentation using Cascaded U-Net & Overall...IRJET-  	  Jeevn-Net: Brain Tumor Segmentation using Cascaded U-Net & Overall...
IRJET- Jeevn-Net: Brain Tumor Segmentation using Cascaded U-Net & Overall...
 
Parallel knn on gpu architecture using opencl
Parallel knn on gpu architecture using openclParallel knn on gpu architecture using opencl
Parallel knn on gpu architecture using opencl
 
Parallel k nn on gpu architecture using opencl
Parallel k nn on gpu architecture using openclParallel k nn on gpu architecture using opencl
Parallel k nn on gpu architecture using opencl
 
Consequence driven decomposition of large-scale power system security analysis
Consequence driven decomposition of large-scale power system security analysisConsequence driven decomposition of large-scale power system security analysis
Consequence driven decomposition of large-scale power system security analysis
 
Threshold adaptation and XOR accumulation algorithm for objects detection
Threshold adaptation and XOR accumulation algorithm for  objects detectionThreshold adaptation and XOR accumulation algorithm for  objects detection
Threshold adaptation and XOR accumulation algorithm for objects detection
 
"An adaptive modular approach to the mining of sensor network ...
"An adaptive modular approach to the mining of sensor network ..."An adaptive modular approach to the mining of sensor network ...
"An adaptive modular approach to the mining of sensor network ...
 
An efficient recovery mechanism
An efficient recovery mechanismAn efficient recovery mechanism
An efficient recovery mechanism
 
Design of Kalman filter for Airborne Applications
Design of Kalman filter for Airborne ApplicationsDesign of Kalman filter for Airborne Applications
Design of Kalman filter for Airborne Applications
 
캡슐 네트워크를 이용한 엔드투엔드 음성 단어 인식, 배재성(KAIST 석사과정)
캡슐 네트워크를 이용한 엔드투엔드 음성 단어 인식, 배재성(KAIST 석사과정)캡슐 네트워크를 이용한 엔드투엔드 음성 단어 인식, 배재성(KAIST 석사과정)
캡슐 네트워크를 이용한 엔드투엔드 음성 단어 인식, 배재성(KAIST 석사과정)
 
Field-programmable gate array design of image encryption and decryption usin...
Field-programmable gate array design of image encryption and  decryption usin...Field-programmable gate array design of image encryption and  decryption usin...
Field-programmable gate array design of image encryption and decryption usin...
 
Deep learning for_ecg_classification
Deep learning for_ecg_classificationDeep learning for_ecg_classification
Deep learning for_ecg_classification
 
1st review android malware.pptx
1st review  android malware.pptx1st review  android malware.pptx
1st review android malware.pptx
 

Recently uploaded

KubeKraft presentation @CloudNativeHooghly
KubeKraft presentation @CloudNativeHooghlyKubeKraft presentation @CloudNativeHooghly
KubeKraft presentation @CloudNativeHooghlysanyuktamishra911
 
AKTU Computer Networks notes --- Unit 3.pdf
AKTU Computer Networks notes ---  Unit 3.pdfAKTU Computer Networks notes ---  Unit 3.pdf
AKTU Computer Networks notes --- Unit 3.pdfankushspencer015
 
Call for Papers - African Journal of Biological Sciences, E-ISSN: 2663-2187, ...
Call for Papers - African Journal of Biological Sciences, E-ISSN: 2663-2187, ...Call for Papers - African Journal of Biological Sciences, E-ISSN: 2663-2187, ...
Call for Papers - African Journal of Biological Sciences, E-ISSN: 2663-2187, ...Christo Ananth
 
Thermal Engineering-R & A / C - unit - V
Thermal Engineering-R & A / C - unit - VThermal Engineering-R & A / C - unit - V
Thermal Engineering-R & A / C - unit - VDineshKumar4165
 
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXssuser89054b
 
Call Girls In Bangalore ☎ 7737669865 🥵 Book Your One night Stand
Call Girls In Bangalore ☎ 7737669865 🥵 Book Your One night StandCall Girls In Bangalore ☎ 7737669865 🥵 Book Your One night Stand
Call Girls In Bangalore ☎ 7737669865 🥵 Book Your One night Standamitlee9823
 
Call for Papers - Educational Administration: Theory and Practice, E-ISSN: 21...
Call for Papers - Educational Administration: Theory and Practice, E-ISSN: 21...Call for Papers - Educational Administration: Theory and Practice, E-ISSN: 21...
Call for Papers - Educational Administration: Theory and Practice, E-ISSN: 21...Christo Ananth
 
VIP Call Girls Palanpur 7001035870 Whatsapp Number, 24/07 Booking
VIP Call Girls Palanpur 7001035870 Whatsapp Number, 24/07 BookingVIP Call Girls Palanpur 7001035870 Whatsapp Number, 24/07 Booking
VIP Call Girls Palanpur 7001035870 Whatsapp Number, 24/07 Bookingdharasingh5698
 
ONLINE FOOD ORDER SYSTEM PROJECT REPORT.pdf
ONLINE FOOD ORDER SYSTEM PROJECT REPORT.pdfONLINE FOOD ORDER SYSTEM PROJECT REPORT.pdf
ONLINE FOOD ORDER SYSTEM PROJECT REPORT.pdfKamal Acharya
 
data_management_and _data_science_cheat_sheet.pdf
data_management_and _data_science_cheat_sheet.pdfdata_management_and _data_science_cheat_sheet.pdf
data_management_and _data_science_cheat_sheet.pdfJiananWang21
 
FULL ENJOY Call Girls In Mahipalpur Delhi Contact Us 8377877756
FULL ENJOY Call Girls In Mahipalpur Delhi Contact Us 8377877756FULL ENJOY Call Girls In Mahipalpur Delhi Contact Us 8377877756
FULL ENJOY Call Girls In Mahipalpur Delhi Contact Us 8377877756dollysharma2066
 
notes on Evolution Of Analytic Scalability.ppt
notes on Evolution Of Analytic Scalability.pptnotes on Evolution Of Analytic Scalability.ppt
notes on Evolution Of Analytic Scalability.pptMsecMca
 
Bhosari ( Call Girls ) Pune 6297143586 Hot Model With Sexy Bhabi Ready For ...
Bhosari ( Call Girls ) Pune  6297143586  Hot Model With Sexy Bhabi Ready For ...Bhosari ( Call Girls ) Pune  6297143586  Hot Model With Sexy Bhabi Ready For ...
Bhosari ( Call Girls ) Pune 6297143586 Hot Model With Sexy Bhabi Ready For ...tanu pandey
 
Online banking management system project.pdf
Online banking management system project.pdfOnline banking management system project.pdf
Online banking management system project.pdfKamal Acharya
 
Call Girls Pimpri Chinchwad Call Me 7737669865 Budget Friendly No Advance Boo...
Call Girls Pimpri Chinchwad Call Me 7737669865 Budget Friendly No Advance Boo...Call Girls Pimpri Chinchwad Call Me 7737669865 Budget Friendly No Advance Boo...
Call Girls Pimpri Chinchwad Call Me 7737669865 Budget Friendly No Advance Boo...roncy bisnoi
 
PVC VS. FIBERGLASS (FRP) GRAVITY SEWER - UNI BELL
PVC VS. FIBERGLASS (FRP) GRAVITY SEWER - UNI BELLPVC VS. FIBERGLASS (FRP) GRAVITY SEWER - UNI BELL
PVC VS. FIBERGLASS (FRP) GRAVITY SEWER - UNI BELLManishPatel169454
 
Call Girls Walvekar Nagar Call Me 7737669865 Budget Friendly No Advance Booking
Call Girls Walvekar Nagar Call Me 7737669865 Budget Friendly No Advance BookingCall Girls Walvekar Nagar Call Me 7737669865 Budget Friendly No Advance Booking
Call Girls Walvekar Nagar Call Me 7737669865 Budget Friendly No Advance Bookingroncy bisnoi
 
Unleashing the Power of the SORA AI lastest leap
Unleashing the Power of the SORA AI lastest leapUnleashing the Power of the SORA AI lastest leap
Unleashing the Power of the SORA AI lastest leapRishantSharmaFr
 

Recently uploaded (20)

NFPA 5000 2024 standard .
NFPA 5000 2024 standard                                  .NFPA 5000 2024 standard                                  .
NFPA 5000 2024 standard .
 
KubeKraft presentation @CloudNativeHooghly
KubeKraft presentation @CloudNativeHooghlyKubeKraft presentation @CloudNativeHooghly
KubeKraft presentation @CloudNativeHooghly
 
AKTU Computer Networks notes --- Unit 3.pdf
AKTU Computer Networks notes ---  Unit 3.pdfAKTU Computer Networks notes ---  Unit 3.pdf
AKTU Computer Networks notes --- Unit 3.pdf
 
Call for Papers - African Journal of Biological Sciences, E-ISSN: 2663-2187, ...
Call for Papers - African Journal of Biological Sciences, E-ISSN: 2663-2187, ...Call for Papers - African Journal of Biological Sciences, E-ISSN: 2663-2187, ...
Call for Papers - African Journal of Biological Sciences, E-ISSN: 2663-2187, ...
 
Thermal Engineering-R & A / C - unit - V
Thermal Engineering-R & A / C - unit - VThermal Engineering-R & A / C - unit - V
Thermal Engineering-R & A / C - unit - V
 
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
 
Call Girls In Bangalore ☎ 7737669865 🥵 Book Your One night Stand
Call Girls In Bangalore ☎ 7737669865 🥵 Book Your One night StandCall Girls In Bangalore ☎ 7737669865 🥵 Book Your One night Stand
Call Girls In Bangalore ☎ 7737669865 🥵 Book Your One night Stand
 
Call for Papers - Educational Administration: Theory and Practice, E-ISSN: 21...
Call for Papers - Educational Administration: Theory and Practice, E-ISSN: 21...Call for Papers - Educational Administration: Theory and Practice, E-ISSN: 21...
Call for Papers - Educational Administration: Theory and Practice, E-ISSN: 21...
 
VIP Call Girls Palanpur 7001035870 Whatsapp Number, 24/07 Booking
VIP Call Girls Palanpur 7001035870 Whatsapp Number, 24/07 BookingVIP Call Girls Palanpur 7001035870 Whatsapp Number, 24/07 Booking
VIP Call Girls Palanpur 7001035870 Whatsapp Number, 24/07 Booking
 
ONLINE FOOD ORDER SYSTEM PROJECT REPORT.pdf
ONLINE FOOD ORDER SYSTEM PROJECT REPORT.pdfONLINE FOOD ORDER SYSTEM PROJECT REPORT.pdf
ONLINE FOOD ORDER SYSTEM PROJECT REPORT.pdf
 
data_management_and _data_science_cheat_sheet.pdf
data_management_and _data_science_cheat_sheet.pdfdata_management_and _data_science_cheat_sheet.pdf
data_management_and _data_science_cheat_sheet.pdf
 
(INDIRA) Call Girl Bhosari Call Now 8617697112 Bhosari Escorts 24x7
(INDIRA) Call Girl Bhosari Call Now 8617697112 Bhosari Escorts 24x7(INDIRA) Call Girl Bhosari Call Now 8617697112 Bhosari Escorts 24x7
(INDIRA) Call Girl Bhosari Call Now 8617697112 Bhosari Escorts 24x7
 
FULL ENJOY Call Girls In Mahipalpur Delhi Contact Us 8377877756
FULL ENJOY Call Girls In Mahipalpur Delhi Contact Us 8377877756FULL ENJOY Call Girls In Mahipalpur Delhi Contact Us 8377877756
FULL ENJOY Call Girls In Mahipalpur Delhi Contact Us 8377877756
 
notes on Evolution Of Analytic Scalability.ppt
notes on Evolution Of Analytic Scalability.pptnotes on Evolution Of Analytic Scalability.ppt
notes on Evolution Of Analytic Scalability.ppt
 
Bhosari ( Call Girls ) Pune 6297143586 Hot Model With Sexy Bhabi Ready For ...
Bhosari ( Call Girls ) Pune  6297143586  Hot Model With Sexy Bhabi Ready For ...Bhosari ( Call Girls ) Pune  6297143586  Hot Model With Sexy Bhabi Ready For ...
Bhosari ( Call Girls ) Pune 6297143586 Hot Model With Sexy Bhabi Ready For ...
 
Online banking management system project.pdf
Online banking management system project.pdfOnline banking management system project.pdf
Online banking management system project.pdf
 
Call Girls Pimpri Chinchwad Call Me 7737669865 Budget Friendly No Advance Boo...
Call Girls Pimpri Chinchwad Call Me 7737669865 Budget Friendly No Advance Boo...Call Girls Pimpri Chinchwad Call Me 7737669865 Budget Friendly No Advance Boo...
Call Girls Pimpri Chinchwad Call Me 7737669865 Budget Friendly No Advance Boo...
 
PVC VS. FIBERGLASS (FRP) GRAVITY SEWER - UNI BELL
PVC VS. FIBERGLASS (FRP) GRAVITY SEWER - UNI BELLPVC VS. FIBERGLASS (FRP) GRAVITY SEWER - UNI BELL
PVC VS. FIBERGLASS (FRP) GRAVITY SEWER - UNI BELL
 
Call Girls Walvekar Nagar Call Me 7737669865 Budget Friendly No Advance Booking
Call Girls Walvekar Nagar Call Me 7737669865 Budget Friendly No Advance BookingCall Girls Walvekar Nagar Call Me 7737669865 Budget Friendly No Advance Booking
Call Girls Walvekar Nagar Call Me 7737669865 Budget Friendly No Advance Booking
 
Unleashing the Power of the SORA AI lastest leap
Unleashing the Power of the SORA AI lastest leapUnleashing the Power of the SORA AI lastest leap
Unleashing the Power of the SORA AI lastest leap
 

ICSRS_R038.pptx

  • 1. Javier Fernández1,3 , Irune Agirre3 , Jon Perez-Cerrolaza3 , Francisco J. Cazorla1 , Jaume Abella1,2
  • 3. 3 Main Concepts Artificial Intelligence It has made enormous progress, reaching near- human accuracy in several safety-related tasks. Functional safety standards IEC 61508 IEC 61513 EN 5012X ISO 26262 Example in the Automotive domain
  • 4. 4 Main Concepts Baseline concept Detection of faults at runtime in the Matrix-matrix Multiplication Catalog of diagnostics techniques CUTLASS: High-performance matrix- matrix multiplication Library “On the Safe Deployment of Matrix Multiplication in Massively Parallel Safety-Related Systems” Object detector application based on CNNs (Tiny YOLO-v3) Matrix-matrix Multiplication (MMM) It is the backbone of the Convolutional Neural Networks in terms of execution time: • Sequential implementation: 98,5 % • Vectorized implementation: 87% • CUDA based implementation: 67 % 𝐴11 𝐴12 𝐴21 𝐴22 𝐴31 𝐴32 𝑋 𝐵11 𝐵12 𝐵13 𝐵21 𝐵22 𝐵23 = 𝐶11 𝐶12 𝐶13 𝐶21 𝐶22 𝐶23 𝐶31 𝐶32 𝐶33
  • 6. 6 P R O P O S E D S O LU T I O N (1) (1) (1) 1 Berkeley DeepDrive dataset (https://www.bdd100k.com/)
  • 7. 7 Stage 1 P R O P O S E D S O LU T I O N 1 Berkeley DeepDrive dataset (https://www.bdd100k.com/) (1) (1) (1)
  • 8. Stage 1 P R O P O S E D S O LU T I O N 8 (1) (1) (1) 1 Berkeley DeepDrive dataset (https://www.bdd100k.com/)
  • 9. 9 Stage 1 P R O P O S E D S O LU T I O N (1) (1) (1) 1 Berkeley DeepDrive dataset (https://www.bdd100k.com/)
  • 10. 10 Stage 2 Stage 1 P R O P O S E D S O LU T I O N (1) (1) (1) 1 Berkeley DeepDrive dataset (https://www.bdd100k.com/)
  • 11. 11 Stage 2 Stage 1 11 P R O P O S E D S O LU T I O N (1) (1) (1) 1 Berkeley DeepDrive dataset (https://www.bdd100k.com/)
  • 12. Stage 2 Stage 1 P R O P O S E D S O LU T I O N 1 Berkeley DeepDrive dataset (https://www.bdd100k.com/) (1) (1) (1) 12
  • 13. Stage 2 Stage 1 P R O P O S E D S O LU T I O N (1) (1) (1) 1 Berkeley DeepDrive dataset (https://www.bdd100k.com/) 13
  • 14. Stage 2 Stage 1 Stage 2 Stage 1 P R O P O S E D S O LU T I O N (1) (1) (1) 1 Berkeley DeepDrive dataset (https://www.bdd100k.com/) 14
  • 15. Stage 2 Stage 1 P R O P O S E D S O LU T I O N (1) (1) (1) 1 Berkeley DeepDrive dataset (https://www.bdd100k.com/)
  • 16. Stage 2 Stage 1 P R O P O S E D S O LU T I O N (1) (1) (1) 1 Berkeley DeepDrive dataset (https://www.bdd100k.com/) 16
  • 17. Stage 2 Stage 1 P R O P O S E D S O LU T I O N 17 DC computation per fault source: B. Fault injected at the global memory level: 𝐷𝑒𝑡𝐴= (𝐵1𝑑𝑒𝑡𝐴 † + 𝐵3𝑑𝑒𝑡𝐴 † ) x 𝑁_𝐵𝑅𝑇1+𝐵2𝑑𝑒𝑡𝐴 ∗ +𝐵4𝑑𝑒𝑡𝐴 ∗ 𝐷𝑒𝑡𝐵=(𝐵1𝑑𝑒𝑡𝐵 ⊗ + 𝐵2𝑑𝑒𝑡𝐵 ⊗ ) x 𝑁_𝐵𝐶𝑇1+𝐵3𝑑𝑒𝑡𝐵 △ +𝐵4𝑑𝑒𝑡𝐵 △ 𝐷𝐶 = 𝐷𝑒𝑡𝐴 + 𝐷𝑒𝑡𝐵 𝑀 + 𝑁 𝑥 𝐾 𝑥 𝑑𝑎𝑡𝑎_𝑠𝑖𝑧𝑒 A. Faults injected at the arithmetic level or at the register level: 𝐷𝐶 = 𝑖=1 4 (𝑁𝑏𝑙𝑜𝑐𝑘𝑠𝐵𝑖 × 𝑁det _𝐵𝑖 ) 𝑁𝑓𝑖
  • 18. Stage 3 Stage 2 Stage 1 P R O P O S E D S O LU T I O N 1 Berkeley DeepDrive dataset (https://www.bdd100k.com/) (1) (1) (1) 18
  • 20. Set-up Matrix Multiplication 𝐴11 𝐴12 𝐴21 𝐴22 𝐴31 𝐴32 𝑋 𝐵11 𝐵12 𝐵13 𝐵21 𝐵22 𝐵23 = 𝐶11 𝐶12 𝐶13 𝐶21 𝐶22 𝐶23 𝐶31 𝐶32 𝐶33 𝑀𝑥𝐾 𝑀𝑥𝑁 𝐾𝑥𝑁 Implementation E VA LUAT I O N
  • 21. Set-up Stage 1 s Sensibility to misclassification E VA LUAT I O N 21
  • 22. Stage 2 Set-up Stage 1 Performance impact (without compiler optimization) Performance impact (maximum compiler optimization) E VA LUAT I O N Performance impact: L1 (Minimum): From 1,01 to 1,37 L3 (Maximum): From 1,002 to 1,18 22 Performance impact: L1 (Minimum): From 1,02 to 82,5 L7 (Maximum): From 1,04 to 171,5
  • 23. Stage 2 DC of each layer of Tiny Yolo-v3 E VA LUAT I O N 23 Set-up Stage 1
  • 24. Stage 3 Selective protection Remarks Note that, while such performance impact is high, it could be reduced if diagnostics are just executed once periodically. For example: For the highest diagnostic coverage PI = 3,8x the CNN execution time Process safety time = 100x a single classification task ---------------------------------------------------------------------------------------------------- PI is lower than 5 % E VA LUAT I O N 24 Stage 2 Set-up Stage 1
  • 26. C O N C LU S I O N S Conclusions Conclusions We propose a methodology to selectively protect CNNs deployed on GPUs decomposed into three stages and demonstrate its applicability on a tiny version of an object detector, tiny YOLO-v3. Additionally, we remark: • For this CNN, we observe a higher tendency to misclassify (from 83,4 to 99,6%) in the initial layers (L1-L8). However, the final layers also present lower but still high misclassification rates (from 55,2 to 74,34%). • For the given example, we observe that the lowest performance impact to achieve high, medium, and low DC ranges is 3,8, 3,33, and 2,61, respectively. 26
  • 27. IKERLAN P.º José María Arizmendiarrieta, 2 - 20500 Arrasate-Mondragón T. +34 943712400 F. +34 943796944 THANK YOU
  • 28. IKERLAN P.º José María Arizmendiarrieta, 2 - 20500 Arrasate-Mondragón T. +34 943712400 F. +34 943796944 NAME: JAVIER FERNÁNDEZ MUÑOZ EMAIL: JAVIER.FERNANDEZ@IKERLAN.ES Acknowledgements: • Ikerlan authors have received funding from the Elkartek grant project KK- 2021/00123 of the Basque government. • BSC authors have been partially supported by the Spanish Ministry of Science and Innovation under grant PID2019- 107255GBC21/AEI/10.13039/501100011033
  • 29. Classification is correct if: 1. The central point of the box is less than 50 pixels away 2. Width and height of the boxes vary by less than 25 pixels 3. Accuracy differs by less than 15%.

Editor's Notes

  1. To employ checksums algorithms as diagnostic techniques to compute an Execution Signature (ES) of all the values of the input and output matrices.
  2. 9.06e9 bits, 13,163 years. 45 ms
  3. evaluates the execution time penalty incurred by each diagnostic technique included in the safe catalog. To this end, we apply the diagnostics in all CNN layers and measure the execution time of each one. This process is repeated for the different types of protection techniques provided in the diagnostics catalog.
  4. computes an array of golden ESs by including the safe library of diagnostic techniques in the MMM execution of each layer (without fault injections).
  5. However, an exhaustive fault injection campaign may be unaffordable for large matrices due to the required number of iterations to cover all input combinations
  6. We denote as B1 those blocks whose dimensions match the size of the blocks launched to the GPU, B2 as those with an equal number of columns but different rows, B3 if the number of rows matches but columns differ, and finally, B4 if both the number of rows and columns differ. In this case, the errors injected in the input matrices A and B affect several blocks. Therefore, a proper DC computation requires verifying if previous blocks have already counted the detected errors. To do this, we propose distinguishing between errors detected from the fault injection in A (DetA) and B (DetB) matrices
  7. FP: the average of new objects that appear or False Positives FN: undetected objects or False Negatives Note that L11 errors do not produce as many FNs and FPs as the rest of the final layers since the concatenation with L5 and the absence of errors on the other branch (L9 and L10) mitigate their appearance.
  8. relative performance impact is quite insensitive to layer dimensions. This increase is associated with the high optimization of the MMM on GPUs. Including a new data (array of ESs) in the computation exacerbates one of the main problems associated with GPU platforms, the bottleneck created for data access. This bottleneck is the main reason for the high-performance impact of the CRC implementation since this diagnostic is based on memory access. Moreover, Fletcher diagnostic has a similar performance to CRC. However, a key reason for this timing penalty lies in using the modulo operator, which is highly inefficient in GPU implementations.
  9. After defining classification features such as the confidence level, error margin, and the total number of possible errors in the weights, we compute a statistically representative random sampling size.
  10. Diagnostic Test Interval: Defined at design, it is the interval between online tests to detect faults in a safety-related system that have a specified diagnostic coverage