Improving the Robustness of Model Compression by On-Manifold Adversarial Training

Despite the advance in deep learning technology, assuring the robustness of deep neural networks (DNNs) is challenging and necessary in safety-critical environments, including automobiles, IoT devices in smart factories, and medical devices, to name a few. Furthermore, recent developments allow us t...
Ausführliche Beschreibung

Gespeichert in:

Autor*in:	Junhyung Kwon [verfasserIn] Sangkyun Lee [verfasserIn]

Format:	E-Artikel
Sprache:	Englisch

Erschienen:	2021

Schlagwörter:	model compression adversarial robustness robust compression on-manifold perturbation

Übergeordnetes Werk:	In: Future Internet - MDPI AG, 2010, 13(2021), 12, p 300
Übergeordnetes Werk:	volume:13 ; year:2021 ; number:12, p 300

Links:	Link aufrufen Link aufrufen Link aufrufen Journal toc

DOI / URN:	10.3390/fi13120300

Katalog-ID:	DOAJ018814255

Internformat


LEADER	01000caa a22002652 4500
001	DOAJ018814255
003	DE-627
005	20240414225247.0
007	cr uuu---uuuuu
008	230226s2021 xx \|\|\|\|\|o 00\| \|\|eng c
024	7		\|a 10.3390/fi13120300 \|2 doi
035			\|a (DE-627)DOAJ018814255
035			\|a (DE-599)DOAJeefa9c16defa469bb5e7866b8829d3b6
040			\|a DE-627 \|b ger \|c DE-627 \|e rakwb
041			\|a eng
050		0	\|a T58.5-58.64
100	0		\|a Junhyung Kwon \|e verfasserin \|4 aut
245	1	0	\|a Improving the Robustness of Model Compression by On-Manifold Adversarial Training
264		1	\|c 2021
336			\|a Text \|b txt \|2 rdacontent
337			\|a Computermedien \|b c \|2 rdamedia
338			\|a Online-Ressource \|b cr \|2 rdacarrier
520			\|a Despite the advance in deep learning technology, assuring the robustness of deep neural networks (DNNs) is challenging and necessary in safety-critical environments, including automobiles, IoT devices in smart factories, and medical devices, to name a few. Furthermore, recent developments allow us to compress DNNs to reduce the size and computational requirements of DNNs to fit them into small embedded devices. However, how robust a compressed DNN can be has not been well studied in addressing its relationship to other critical factors, such as prediction performance and model sizes. In particular, existing studies on robust model compression have been focused on the robustness against off-manifold adversarial perturbation, which does not explain how a DNN will behave against perturbations that follow the same probability distribution as the training data. This aspect is relevant for on-device AI models, which are more likely to experience perturbations due to noise from the regular data observation environment compared with off-manifold perturbations provided by an external attacker. Therefore, this paper investigates the robustness of compressed deep neural networks, focusing on the relationship between the model sizes and the prediction performance on noisy perturbations. Our experiment shows that on-manifold adversarial training can be effective in building robust classifiers, especially when the model compression rate is high.
650		4	\|a model compression
650		4	\|a adversarial robustness
650		4	\|a robust compression
650		4	\|a on-manifold perturbation
653		0	\|a Information technology
700	0		\|a Sangkyun Lee \|e verfasserin \|4 aut
773	0	8	\|i In \|t Future Internet \|d MDPI AG, 2010 \|g 13(2021), 12, p 300 \|w (DE-627)610604147 \|w (DE-600)2518385-0 \|x 19995903 \|7 nnns
773	1	8	\|g volume:13 \|g year:2021 \|g number:12, p 300
856	4	0	\|u https://doi.org/10.3390/fi13120300 \|z kostenfrei
856	4	0	\|u https://doaj.org/article/eefa9c16defa469bb5e7866b8829d3b6 \|z kostenfrei
856	4	0	\|u https://www.mdpi.com/1999-5903/13/12/300 \|z kostenfrei
856	4	2	\|u https://doaj.org/toc/1999-5903 \|y Journal toc \|z kostenfrei
912			\|a GBV_USEFLAG_A
912			\|a SYSFLAG_A
912			\|a GBV_DOAJ
912			\|a GBV_ILN_11
912			\|a GBV_ILN_20
912			\|a GBV_ILN_22
912			\|a GBV_ILN_23
912			\|a GBV_ILN_24
912			\|a GBV_ILN_39
912			\|a GBV_ILN_40
912			\|a GBV_ILN_60
912			\|a GBV_ILN_62
912			\|a GBV_ILN_63
912			\|a GBV_ILN_65
912			\|a GBV_ILN_69
912			\|a GBV_ILN_70
912			\|a GBV_ILN_73
912			\|a GBV_ILN_95
912			\|a GBV_ILN_105
912			\|a GBV_ILN_110
912			\|a GBV_ILN_151
912			\|a GBV_ILN_161
912			\|a GBV_ILN_170
912			\|a GBV_ILN_213
912			\|a GBV_ILN_230
912			\|a GBV_ILN_285
912			\|a GBV_ILN_293
912			\|a GBV_ILN_370
912			\|a GBV_ILN_602
912			\|a GBV_ILN_2009
912			\|a GBV_ILN_2014
912			\|a GBV_ILN_2111
912			\|a GBV_ILN_2129
912			\|a GBV_ILN_4012
912			\|a GBV_ILN_4037
912			\|a GBV_ILN_4046
912			\|a GBV_ILN_4112
912			\|a GBV_ILN_4125
912			\|a GBV_ILN_4126
912			\|a GBV_ILN_4249
912			\|a GBV_ILN_4305
912			\|a GBV_ILN_4306
912			\|a GBV_ILN_4307
912			\|a GBV_ILN_4313
912			\|a GBV_ILN_4322
912			\|a GBV_ILN_4323
912			\|a GBV_ILN_4324
912			\|a GBV_ILN_4325
912			\|a GBV_ILN_4326
912			\|a GBV_ILN_4335
912			\|a GBV_ILN_4338
912			\|a GBV_ILN_4367
912			\|a GBV_ILN_4700
951			\|a AR
952			\|d 13 \|j 2021 \|e 12, p 300

Indexfelder

author_variant	j k jk s l sl
matchkey_str	article:19995903:2021----::mrvnteoutesfoecmrsinynaio
hierarchy_sort_str	2021
callnumber-subject-code	T
publishDate	2021
allfields	10.3390/fi13120300 doi (DE-627)DOAJ018814255 (DE-599)DOAJeefa9c16defa469bb5e7866b8829d3b6 DE-627 ger DE-627 rakwb eng T58.5-58.64 Junhyung Kwon verfasserin aut Improving the Robustness of Model Compression by On-Manifold Adversarial Training 2021 Text txt rdacontent Computermedien c rdamedia Online-Ressource cr rdacarrier Despite the advance in deep learning technology, assuring the robustness of deep neural networks (DNNs) is challenging and necessary in safety-critical environments, including automobiles, IoT devices in smart factories, and medical devices, to name a few. Furthermore, recent developments allow us to compress DNNs to reduce the size and computational requirements of DNNs to fit them into small embedded devices. However, how robust a compressed DNN can be has not been well studied in addressing its relationship to other critical factors, such as prediction performance and model sizes. In particular, existing studies on robust model compression have been focused on the robustness against off-manifold adversarial perturbation, which does not explain how a DNN will behave against perturbations that follow the same probability distribution as the training data. This aspect is relevant for on-device AI models, which are more likely to experience perturbations due to noise from the regular data observation environment compared with off-manifold perturbations provided by an external attacker. Therefore, this paper investigates the robustness of compressed deep neural networks, focusing on the relationship between the model sizes and the prediction performance on noisy perturbations. Our experiment shows that on-manifold adversarial training can be effective in building robust classifiers, especially when the model compression rate is high. model compression adversarial robustness robust compression on-manifold perturbation Information technology Sangkyun Lee verfasserin aut In Future Internet MDPI AG, 2010 13(2021), 12, p 300 (DE-627)610604147 (DE-600)2518385-0 19995903 nnns volume:13 year:2021 number:12, p 300 https://doi.org/10.3390/fi13120300 kostenfrei https://doaj.org/article/eefa9c16defa469bb5e7866b8829d3b6 kostenfrei https://www.mdpi.com/1999-5903/13/12/300 kostenfrei https://doaj.org/toc/1999-5903 Journal toc kostenfrei GBV_USEFLAG_A SYSFLAG_A GBV_DOAJ GBV_ILN_11 GBV_ILN_20 GBV_ILN_22 GBV_ILN_23 GBV_ILN_24 GBV_ILN_39 GBV_ILN_40 GBV_ILN_60 GBV_ILN_62 GBV_ILN_63 GBV_ILN_65 GBV_ILN_69 GBV_ILN_70 GBV_ILN_73 GBV_ILN_95 GBV_ILN_105 GBV_ILN_110 GBV_ILN_151 GBV_ILN_161 GBV_ILN_170 GBV_ILN_213 GBV_ILN_230 GBV_ILN_285 GBV_ILN_293 GBV_ILN_370 GBV_ILN_602 GBV_ILN_2009 GBV_ILN_2014 GBV_ILN_2111 GBV_ILN_2129 GBV_ILN_4012 GBV_ILN_4037 GBV_ILN_4046 GBV_ILN_4112 GBV_ILN_4125 GBV_ILN_4126 GBV_ILN_4249 GBV_ILN_4305 GBV_ILN_4306 GBV_ILN_4307 GBV_ILN_4313 GBV_ILN_4322 GBV_ILN_4323 GBV_ILN_4324 GBV_ILN_4325 GBV_ILN_4326 GBV_ILN_4335 GBV_ILN_4338 GBV_ILN_4367 GBV_ILN_4700 AR 13 2021 12, p 300
spelling	10.3390/fi13120300 doi (DE-627)DOAJ018814255 (DE-599)DOAJeefa9c16defa469bb5e7866b8829d3b6 DE-627 ger DE-627 rakwb eng T58.5-58.64 Junhyung Kwon verfasserin aut Improving the Robustness of Model Compression by On-Manifold Adversarial Training 2021 Text txt rdacontent Computermedien c rdamedia Online-Ressource cr rdacarrier Despite the advance in deep learning technology, assuring the robustness of deep neural networks (DNNs) is challenging and necessary in safety-critical environments, including automobiles, IoT devices in smart factories, and medical devices, to name a few. Furthermore, recent developments allow us to compress DNNs to reduce the size and computational requirements of DNNs to fit them into small embedded devices. However, how robust a compressed DNN can be has not been well studied in addressing its relationship to other critical factors, such as prediction performance and model sizes. In particular, existing studies on robust model compression have been focused on the robustness against off-manifold adversarial perturbation, which does not explain how a DNN will behave against perturbations that follow the same probability distribution as the training data. This aspect is relevant for on-device AI models, which are more likely to experience perturbations due to noise from the regular data observation environment compared with off-manifold perturbations provided by an external attacker. Therefore, this paper investigates the robustness of compressed deep neural networks, focusing on the relationship between the model sizes and the prediction performance on noisy perturbations. Our experiment shows that on-manifold adversarial training can be effective in building robust classifiers, especially when the model compression rate is high. model compression adversarial robustness robust compression on-manifold perturbation Information technology Sangkyun Lee verfasserin aut In Future Internet MDPI AG, 2010 13(2021), 12, p 300 (DE-627)610604147 (DE-600)2518385-0 19995903 nnns volume:13 year:2021 number:12, p 300 https://doi.org/10.3390/fi13120300 kostenfrei https://doaj.org/article/eefa9c16defa469bb5e7866b8829d3b6 kostenfrei https://www.mdpi.com/1999-5903/13/12/300 kostenfrei https://doaj.org/toc/1999-5903 Journal toc kostenfrei GBV_USEFLAG_A SYSFLAG_A GBV_DOAJ GBV_ILN_11 GBV_ILN_20 GBV_ILN_22 GBV_ILN_23 GBV_ILN_24 GBV_ILN_39 GBV_ILN_40 GBV_ILN_60 GBV_ILN_62 GBV_ILN_63 GBV_ILN_65 GBV_ILN_69 GBV_ILN_70 GBV_ILN_73 GBV_ILN_95 GBV_ILN_105 GBV_ILN_110 GBV_ILN_151 GBV_ILN_161 GBV_ILN_170 GBV_ILN_213 GBV_ILN_230 GBV_ILN_285 GBV_ILN_293 GBV_ILN_370 GBV_ILN_602 GBV_ILN_2009 GBV_ILN_2014 GBV_ILN_2111 GBV_ILN_2129 GBV_ILN_4012 GBV_ILN_4037 GBV_ILN_4046 GBV_ILN_4112 GBV_ILN_4125 GBV_ILN_4126 GBV_ILN_4249 GBV_ILN_4305 GBV_ILN_4306 GBV_ILN_4307 GBV_ILN_4313 GBV_ILN_4322 GBV_ILN_4323 GBV_ILN_4324 GBV_ILN_4325 GBV_ILN_4326 GBV_ILN_4335 GBV_ILN_4338 GBV_ILN_4367 GBV_ILN_4700 AR 13 2021 12, p 300
allfields_unstemmed	10.3390/fi13120300 doi (DE-627)DOAJ018814255 (DE-599)DOAJeefa9c16defa469bb5e7866b8829d3b6 DE-627 ger DE-627 rakwb eng T58.5-58.64 Junhyung Kwon verfasserin aut Improving the Robustness of Model Compression by On-Manifold Adversarial Training 2021 Text txt rdacontent Computermedien c rdamedia Online-Ressource cr rdacarrier Despite the advance in deep learning technology, assuring the robustness of deep neural networks (DNNs) is challenging and necessary in safety-critical environments, including automobiles, IoT devices in smart factories, and medical devices, to name a few. Furthermore, recent developments allow us to compress DNNs to reduce the size and computational requirements of DNNs to fit them into small embedded devices. However, how robust a compressed DNN can be has not been well studied in addressing its relationship to other critical factors, such as prediction performance and model sizes. In particular, existing studies on robust model compression have been focused on the robustness against off-manifold adversarial perturbation, which does not explain how a DNN will behave against perturbations that follow the same probability distribution as the training data. This aspect is relevant for on-device AI models, which are more likely to experience perturbations due to noise from the regular data observation environment compared with off-manifold perturbations provided by an external attacker. Therefore, this paper investigates the robustness of compressed deep neural networks, focusing on the relationship between the model sizes and the prediction performance on noisy perturbations. Our experiment shows that on-manifold adversarial training can be effective in building robust classifiers, especially when the model compression rate is high. model compression adversarial robustness robust compression on-manifold perturbation Information technology Sangkyun Lee verfasserin aut In Future Internet MDPI AG, 2010 13(2021), 12, p 300 (DE-627)610604147 (DE-600)2518385-0 19995903 nnns volume:13 year:2021 number:12, p 300 https://doi.org/10.3390/fi13120300 kostenfrei https://doaj.org/article/eefa9c16defa469bb5e7866b8829d3b6 kostenfrei https://www.mdpi.com/1999-5903/13/12/300 kostenfrei https://doaj.org/toc/1999-5903 Journal toc kostenfrei GBV_USEFLAG_A SYSFLAG_A GBV_DOAJ GBV_ILN_11 GBV_ILN_20 GBV_ILN_22 GBV_ILN_23 GBV_ILN_24 GBV_ILN_39 GBV_ILN_40 GBV_ILN_60 GBV_ILN_62 GBV_ILN_63 GBV_ILN_65 GBV_ILN_69 GBV_ILN_70 GBV_ILN_73 GBV_ILN_95 GBV_ILN_105 GBV_ILN_110 GBV_ILN_151 GBV_ILN_161 GBV_ILN_170 GBV_ILN_213 GBV_ILN_230 GBV_ILN_285 GBV_ILN_293 GBV_ILN_370 GBV_ILN_602 GBV_ILN_2009 GBV_ILN_2014 GBV_ILN_2111 GBV_ILN_2129 GBV_ILN_4012 GBV_ILN_4037 GBV_ILN_4046 GBV_ILN_4112 GBV_ILN_4125 GBV_ILN_4126 GBV_ILN_4249 GBV_ILN_4305 GBV_ILN_4306 GBV_ILN_4307 GBV_ILN_4313 GBV_ILN_4322 GBV_ILN_4323 GBV_ILN_4324 GBV_ILN_4325 GBV_ILN_4326 GBV_ILN_4335 GBV_ILN_4338 GBV_ILN_4367 GBV_ILN_4700 AR 13 2021 12, p 300
allfieldsGer	10.3390/fi13120300 doi (DE-627)DOAJ018814255 (DE-599)DOAJeefa9c16defa469bb5e7866b8829d3b6 DE-627 ger DE-627 rakwb eng T58.5-58.64 Junhyung Kwon verfasserin aut Improving the Robustness of Model Compression by On-Manifold Adversarial Training 2021 Text txt rdacontent Computermedien c rdamedia Online-Ressource cr rdacarrier Despite the advance in deep learning technology, assuring the robustness of deep neural networks (DNNs) is challenging and necessary in safety-critical environments, including automobiles, IoT devices in smart factories, and medical devices, to name a few. Furthermore, recent developments allow us to compress DNNs to reduce the size and computational requirements of DNNs to fit them into small embedded devices. However, how robust a compressed DNN can be has not been well studied in addressing its relationship to other critical factors, such as prediction performance and model sizes. In particular, existing studies on robust model compression have been focused on the robustness against off-manifold adversarial perturbation, which does not explain how a DNN will behave against perturbations that follow the same probability distribution as the training data. This aspect is relevant for on-device AI models, which are more likely to experience perturbations due to noise from the regular data observation environment compared with off-manifold perturbations provided by an external attacker. Therefore, this paper investigates the robustness of compressed deep neural networks, focusing on the relationship between the model sizes and the prediction performance on noisy perturbations. Our experiment shows that on-manifold adversarial training can be effective in building robust classifiers, especially when the model compression rate is high. model compression adversarial robustness robust compression on-manifold perturbation Information technology Sangkyun Lee verfasserin aut In Future Internet MDPI AG, 2010 13(2021), 12, p 300 (DE-627)610604147 (DE-600)2518385-0 19995903 nnns volume:13 year:2021 number:12, p 300 https://doi.org/10.3390/fi13120300 kostenfrei https://doaj.org/article/eefa9c16defa469bb5e7866b8829d3b6 kostenfrei https://www.mdpi.com/1999-5903/13/12/300 kostenfrei https://doaj.org/toc/1999-5903 Journal toc kostenfrei GBV_USEFLAG_A SYSFLAG_A GBV_DOAJ GBV_ILN_11 GBV_ILN_20 GBV_ILN_22 GBV_ILN_23 GBV_ILN_24 GBV_ILN_39 GBV_ILN_40 GBV_ILN_60 GBV_ILN_62 GBV_ILN_63 GBV_ILN_65 GBV_ILN_69 GBV_ILN_70 GBV_ILN_73 GBV_ILN_95 GBV_ILN_105 GBV_ILN_110 GBV_ILN_151 GBV_ILN_161 GBV_ILN_170 GBV_ILN_213 GBV_ILN_230 GBV_ILN_285 GBV_ILN_293 GBV_ILN_370 GBV_ILN_602 GBV_ILN_2009 GBV_ILN_2014 GBV_ILN_2111 GBV_ILN_2129 GBV_ILN_4012 GBV_ILN_4037 GBV_ILN_4046 GBV_ILN_4112 GBV_ILN_4125 GBV_ILN_4126 GBV_ILN_4249 GBV_ILN_4305 GBV_ILN_4306 GBV_ILN_4307 GBV_ILN_4313 GBV_ILN_4322 GBV_ILN_4323 GBV_ILN_4324 GBV_ILN_4325 GBV_ILN_4326 GBV_ILN_4335 GBV_ILN_4338 GBV_ILN_4367 GBV_ILN_4700 AR 13 2021 12, p 300
allfieldsSound	10.3390/fi13120300 doi (DE-627)DOAJ018814255 (DE-599)DOAJeefa9c16defa469bb5e7866b8829d3b6 DE-627 ger DE-627 rakwb eng T58.5-58.64 Junhyung Kwon verfasserin aut Improving the Robustness of Model Compression by On-Manifold Adversarial Training 2021 Text txt rdacontent Computermedien c rdamedia Online-Ressource cr rdacarrier Despite the advance in deep learning technology, assuring the robustness of deep neural networks (DNNs) is challenging and necessary in safety-critical environments, including automobiles, IoT devices in smart factories, and medical devices, to name a few. Furthermore, recent developments allow us to compress DNNs to reduce the size and computational requirements of DNNs to fit them into small embedded devices. However, how robust a compressed DNN can be has not been well studied in addressing its relationship to other critical factors, such as prediction performance and model sizes. In particular, existing studies on robust model compression have been focused on the robustness against off-manifold adversarial perturbation, which does not explain how a DNN will behave against perturbations that follow the same probability distribution as the training data. This aspect is relevant for on-device AI models, which are more likely to experience perturbations due to noise from the regular data observation environment compared with off-manifold perturbations provided by an external attacker. Therefore, this paper investigates the robustness of compressed deep neural networks, focusing on the relationship between the model sizes and the prediction performance on noisy perturbations. Our experiment shows that on-manifold adversarial training can be effective in building robust classifiers, especially when the model compression rate is high. model compression adversarial robustness robust compression on-manifold perturbation Information technology Sangkyun Lee verfasserin aut In Future Internet MDPI AG, 2010 13(2021), 12, p 300 (DE-627)610604147 (DE-600)2518385-0 19995903 nnns volume:13 year:2021 number:12, p 300 https://doi.org/10.3390/fi13120300 kostenfrei https://doaj.org/article/eefa9c16defa469bb5e7866b8829d3b6 kostenfrei https://www.mdpi.com/1999-5903/13/12/300 kostenfrei https://doaj.org/toc/1999-5903 Journal toc kostenfrei GBV_USEFLAG_A SYSFLAG_A GBV_DOAJ GBV_ILN_11 GBV_ILN_20 GBV_ILN_22 GBV_ILN_23 GBV_ILN_24 GBV_ILN_39 GBV_ILN_40 GBV_ILN_60 GBV_ILN_62 GBV_ILN_63 GBV_ILN_65 GBV_ILN_69 GBV_ILN_70 GBV_ILN_73 GBV_ILN_95 GBV_ILN_105 GBV_ILN_110 GBV_ILN_151 GBV_ILN_161 GBV_ILN_170 GBV_ILN_213 GBV_ILN_230 GBV_ILN_285 GBV_ILN_293 GBV_ILN_370 GBV_ILN_602 GBV_ILN_2009 GBV_ILN_2014 GBV_ILN_2111 GBV_ILN_2129 GBV_ILN_4012 GBV_ILN_4037 GBV_ILN_4046 GBV_ILN_4112 GBV_ILN_4125 GBV_ILN_4126 GBV_ILN_4249 GBV_ILN_4305 GBV_ILN_4306 GBV_ILN_4307 GBV_ILN_4313 GBV_ILN_4322 GBV_ILN_4323 GBV_ILN_4324 GBV_ILN_4325 GBV_ILN_4326 GBV_ILN_4335 GBV_ILN_4338 GBV_ILN_4367 GBV_ILN_4700 AR 13 2021 12, p 300
language	English
source	In Future Internet 13(2021), 12, p 300 volume:13 year:2021 number:12, p 300
sourceStr	In Future Internet 13(2021), 12, p 300 volume:13 year:2021 number:12, p 300
format_phy_str_mv	Article
institution	findex.gbv.de
topic_facet	model compression adversarial robustness robust compression on-manifold perturbation Information technology
isfreeaccess_bool	true
container_title	Future Internet
authorswithroles_txt_mv	Junhyung Kwon @@aut@@ Sangkyun Lee @@aut@@
publishDateDaySort_date	2021-01-01T00:00:00Z
hierarchy_top_id	610604147
id	DOAJ018814255
language_de	englisch
fullrecord	<?xml version="1.0" encoding="UTF-8"?><collection xmlns="http://www.loc.gov/MARC21/slim"><record><leader>01000caa a22002652 4500</leader><controlfield tag="001">DOAJ018814255</controlfield><controlfield tag="003">DE-627</controlfield><controlfield tag="005">20240414225247.0</controlfield><controlfield tag="007">cr uuu---uuuuu</controlfield><controlfield tag="008">230226s2021 xx \|\|\|\|\|o 00\| \|\|eng c</controlfield><datafield tag="024" ind1="7" ind2=" "><subfield code="a">10.3390/fi13120300</subfield><subfield code="2">doi</subfield></datafield><datafield tag="035" ind1=" " ind2=" "><subfield code="a">(DE-627)DOAJ018814255</subfield></datafield><datafield tag="035" ind1=" " ind2=" "><subfield code="a">(DE-599)DOAJeefa9c16defa469bb5e7866b8829d3b6</subfield></datafield><datafield tag="040" ind1=" " ind2=" "><subfield code="a">DE-627</subfield><subfield code="b">ger</subfield><subfield code="c">DE-627</subfield><subfield code="e">rakwb</subfield></datafield><datafield tag="041" ind1=" " ind2=" "><subfield code="a">eng</subfield></datafield><datafield tag="050" ind1=" " ind2="0"><subfield code="a">T58.5-58.64</subfield></datafield><datafield tag="100" ind1="0" ind2=" "><subfield code="a">Junhyung Kwon</subfield><subfield code="e">verfasserin</subfield><subfield code="4">aut</subfield></datafield><datafield tag="245" ind1="1" ind2="0"><subfield code="a">Improving the Robustness of Model Compression by On-Manifold Adversarial Training</subfield></datafield><datafield tag="264" ind1=" " ind2="1"><subfield code="c">2021</subfield></datafield><datafield tag="336" ind1=" " ind2=" "><subfield code="a">Text</subfield><subfield code="b">txt</subfield><subfield code="2">rdacontent</subfield></datafield><datafield tag="337" ind1=" " ind2=" "><subfield code="a">Computermedien</subfield><subfield code="b">c</subfield><subfield code="2">rdamedia</subfield></datafield><datafield tag="338" ind1=" " ind2=" "><subfield code="a">Online-Ressource</subfield><subfield code="b">cr</subfield><subfield code="2">rdacarrier</subfield></datafield><datafield tag="520" ind1=" " ind2=" "><subfield code="a">Despite the advance in deep learning technology, assuring the robustness of deep neural networks (DNNs) is challenging and necessary in safety-critical environments, including automobiles, IoT devices in smart factories, and medical devices, to name a few. Furthermore, recent developments allow us to compress DNNs to reduce the size and computational requirements of DNNs to fit them into small embedded devices. However, how robust a compressed DNN can be has not been well studied in addressing its relationship to other critical factors, such as prediction performance and model sizes. In particular, existing studies on robust model compression have been focused on the robustness against off-manifold adversarial perturbation, which does not explain how a DNN will behave against perturbations that follow the same probability distribution as the training data. This aspect is relevant for on-device AI models, which are more likely to experience perturbations due to noise from the regular data observation environment compared with off-manifold perturbations provided by an external attacker. Therefore, this paper investigates the robustness of compressed deep neural networks, focusing on the relationship between the model sizes and the prediction performance on noisy perturbations. Our experiment shows that on-manifold adversarial training can be effective in building robust classifiers, especially when the model compression rate is high.</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">model compression</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">adversarial robustness</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">robust compression</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">on-manifold perturbation</subfield></datafield><datafield tag="653" ind1=" " ind2="0"><subfield code="a">Information technology</subfield></datafield><datafield tag="700" ind1="0" ind2=" "><subfield code="a">Sangkyun Lee</subfield><subfield code="e">verfasserin</subfield><subfield code="4">aut</subfield></datafield><datafield tag="773" ind1="0" ind2="8"><subfield code="i">In</subfield><subfield code="t">Future Internet</subfield><subfield code="d">MDPI AG, 2010</subfield><subfield code="g">13(2021), 12, p 300</subfield><subfield code="w">(DE-627)610604147</subfield><subfield code="w">(DE-600)2518385-0</subfield><subfield code="x">19995903</subfield><subfield code="7">nnns</subfield></datafield><datafield tag="773" ind1="1" ind2="8"><subfield code="g">volume:13</subfield><subfield code="g">year:2021</subfield><subfield code="g">number:12, p 300</subfield></datafield><datafield tag="856" ind1="4" ind2="0"><subfield code="u">https://doi.org/10.3390/fi13120300</subfield><subfield code="z">kostenfrei</subfield></datafield><datafield tag="856" ind1="4" ind2="0"><subfield code="u">https://doaj.org/article/eefa9c16defa469bb5e7866b8829d3b6</subfield><subfield code="z">kostenfrei</subfield></datafield><datafield tag="856" ind1="4" ind2="0"><subfield code="u">https://www.mdpi.com/1999-5903/13/12/300</subfield><subfield code="z">kostenfrei</subfield></datafield><datafield tag="856" ind1="4" ind2="2"><subfield code="u">https://doaj.org/toc/1999-5903</subfield><subfield code="y">Journal toc</subfield><subfield code="z">kostenfrei</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_USEFLAG_A</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">SYSFLAG_A</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_DOAJ</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_11</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_20</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_22</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_23</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_24</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_39</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_40</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_60</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_62</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_63</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_65</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_69</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_70</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_73</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_95</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_105</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_110</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_151</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_161</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_170</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_213</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_230</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_285</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_293</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_370</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_602</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_2009</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_2014</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_2111</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_2129</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4012</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4037</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4046</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4112</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4125</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4126</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4249</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4305</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4306</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4307</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4313</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4322</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4323</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4324</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4325</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4326</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4335</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4338</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4367</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4700</subfield></datafield><datafield tag="951" ind1=" " ind2=" "><subfield code="a">AR</subfield></datafield><datafield tag="952" ind1=" " ind2=" "><subfield code="d">13</subfield><subfield code="j">2021</subfield><subfield code="e">12, p 300</subfield></datafield></record></collection>
callnumber-first	T - Technology
author	Junhyung Kwon
spellingShingle	Junhyung Kwon misc T58.5-58.64 misc model compression misc adversarial robustness misc robust compression misc on-manifold perturbation misc Information technology Improving the Robustness of Model Compression by On-Manifold Adversarial Training
authorStr	Junhyung Kwon
ppnlink_with_tag_str_mv	@@773@@(DE-627)610604147
format	electronic Article
delete_txt_mv	keep
author_role	aut aut
collection	DOAJ
remote_str	true
callnumber-label	T58
illustrated	Not Illustrated
issn	19995903
topic_title	T58.5-58.64 Improving the Robustness of Model Compression by On-Manifold Adversarial Training model compression adversarial robustness robust compression on-manifold perturbation
topic	misc T58.5-58.64 misc model compression misc adversarial robustness misc robust compression misc on-manifold perturbation misc Information technology
topic_unstemmed	misc T58.5-58.64 misc model compression misc adversarial robustness misc robust compression misc on-manifold perturbation misc Information technology
topic_browse	misc T58.5-58.64 misc model compression misc adversarial robustness misc robust compression misc on-manifold perturbation misc Information technology
format_facet	Elektronische Aufsätze Aufsätze Elektronische Ressource
format_main_str_mv	Text Zeitschrift/Artikel
carriertype_str_mv	cr
hierarchy_parent_title	Future Internet
hierarchy_parent_id	610604147
hierarchy_top_title	Future Internet
isfreeaccess_txt	true
familylinks_str_mv	(DE-627)610604147 (DE-600)2518385-0
title	Improving the Robustness of Model Compression by On-Manifold Adversarial Training
ctrlnum	(DE-627)DOAJ018814255 (DE-599)DOAJeefa9c16defa469bb5e7866b8829d3b6
title_full	Improving the Robustness of Model Compression by On-Manifold Adversarial Training
author_sort	Junhyung Kwon
journal	Future Internet
journalStr	Future Internet
callnumber-first-code	T
lang_code	eng
isOA_bool	true
recordtype	marc
publishDateSort	2021
contenttype_str_mv	txt
author_browse	Junhyung Kwon Sangkyun Lee
container_volume	13
class	T58.5-58.64
format_se	Elektronische Aufsätze
author-letter	Junhyung Kwon
doi_str_mv	10.3390/fi13120300
author2-role	verfasserin
title_sort	improving the robustness of model compression by on-manifold adversarial training
callnumber	T58.5-58.64
title_auth	Improving the Robustness of Model Compression by On-Manifold Adversarial Training
abstract	Despite the advance in deep learning technology, assuring the robustness of deep neural networks (DNNs) is challenging and necessary in safety-critical environments, including automobiles, IoT devices in smart factories, and medical devices, to name a few. Furthermore, recent developments allow us to compress DNNs to reduce the size and computational requirements of DNNs to fit them into small embedded devices. However, how robust a compressed DNN can be has not been well studied in addressing its relationship to other critical factors, such as prediction performance and model sizes. In particular, existing studies on robust model compression have been focused on the robustness against off-manifold adversarial perturbation, which does not explain how a DNN will behave against perturbations that follow the same probability distribution as the training data. This aspect is relevant for on-device AI models, which are more likely to experience perturbations due to noise from the regular data observation environment compared with off-manifold perturbations provided by an external attacker. Therefore, this paper investigates the robustness of compressed deep neural networks, focusing on the relationship between the model sizes and the prediction performance on noisy perturbations. Our experiment shows that on-manifold adversarial training can be effective in building robust classifiers, especially when the model compression rate is high.
abstractGer	Despite the advance in deep learning technology, assuring the robustness of deep neural networks (DNNs) is challenging and necessary in safety-critical environments, including automobiles, IoT devices in smart factories, and medical devices, to name a few. Furthermore, recent developments allow us to compress DNNs to reduce the size and computational requirements of DNNs to fit them into small embedded devices. However, how robust a compressed DNN can be has not been well studied in addressing its relationship to other critical factors, such as prediction performance and model sizes. In particular, existing studies on robust model compression have been focused on the robustness against off-manifold adversarial perturbation, which does not explain how a DNN will behave against perturbations that follow the same probability distribution as the training data. This aspect is relevant for on-device AI models, which are more likely to experience perturbations due to noise from the regular data observation environment compared with off-manifold perturbations provided by an external attacker. Therefore, this paper investigates the robustness of compressed deep neural networks, focusing on the relationship between the model sizes and the prediction performance on noisy perturbations. Our experiment shows that on-manifold adversarial training can be effective in building robust classifiers, especially when the model compression rate is high.
abstract_unstemmed	Despite the advance in deep learning technology, assuring the robustness of deep neural networks (DNNs) is challenging and necessary in safety-critical environments, including automobiles, IoT devices in smart factories, and medical devices, to name a few. Furthermore, recent developments allow us to compress DNNs to reduce the size and computational requirements of DNNs to fit them into small embedded devices. However, how robust a compressed DNN can be has not been well studied in addressing its relationship to other critical factors, such as prediction performance and model sizes. In particular, existing studies on robust model compression have been focused on the robustness against off-manifold adversarial perturbation, which does not explain how a DNN will behave against perturbations that follow the same probability distribution as the training data. This aspect is relevant for on-device AI models, which are more likely to experience perturbations due to noise from the regular data observation environment compared with off-manifold perturbations provided by an external attacker. Therefore, this paper investigates the robustness of compressed deep neural networks, focusing on the relationship between the model sizes and the prediction performance on noisy perturbations. Our experiment shows that on-manifold adversarial training can be effective in building robust classifiers, especially when the model compression rate is high.
collection_details	GBV_USEFLAG_A SYSFLAG_A GBV_DOAJ GBV_ILN_11 GBV_ILN_20 GBV_ILN_22 GBV_ILN_23 GBV_ILN_24 GBV_ILN_39 GBV_ILN_40 GBV_ILN_60 GBV_ILN_62 GBV_ILN_63 GBV_ILN_65 GBV_ILN_69 GBV_ILN_70 GBV_ILN_73 GBV_ILN_95 GBV_ILN_105 GBV_ILN_110 GBV_ILN_151 GBV_ILN_161 GBV_ILN_170 GBV_ILN_213 GBV_ILN_230 GBV_ILN_285 GBV_ILN_293 GBV_ILN_370 GBV_ILN_602 GBV_ILN_2009 GBV_ILN_2014 GBV_ILN_2111 GBV_ILN_2129 GBV_ILN_4012 GBV_ILN_4037 GBV_ILN_4046 GBV_ILN_4112 GBV_ILN_4125 GBV_ILN_4126 GBV_ILN_4249 GBV_ILN_4305 GBV_ILN_4306 GBV_ILN_4307 GBV_ILN_4313 GBV_ILN_4322 GBV_ILN_4323 GBV_ILN_4324 GBV_ILN_4325 GBV_ILN_4326 GBV_ILN_4335 GBV_ILN_4338 GBV_ILN_4367 GBV_ILN_4700
container_issue	12, p 300
title_short	Improving the Robustness of Model Compression by On-Manifold Adversarial Training
url	https://doi.org/10.3390/fi13120300 https://doaj.org/article/eefa9c16defa469bb5e7866b8829d3b6 https://www.mdpi.com/1999-5903/13/12/300 https://doaj.org/toc/1999-5903
remote_bool	true
author2	Sangkyun Lee
author2Str	Sangkyun Lee
ppnlink	610604147
callnumber-subject	T - General Technology
mediatype_str_mv	c
isOA_txt	true
hochschulschrift_bool	false
doi_str	10.3390/fi13120300
callnumber-a	T58.5-58.64
up_date	2024-07-03T20:08:56.327Z
_version_	1803589869982711808
fullrecord_marcxml	<?xml version="1.0" encoding="UTF-8"?><collection xmlns="http://www.loc.gov/MARC21/slim"><record><leader>01000caa a22002652 4500</leader><controlfield tag="001">DOAJ018814255</controlfield><controlfield tag="003">DE-627</controlfield><controlfield tag="005">20240414225247.0</controlfield><controlfield tag="007">cr uuu---uuuuu</controlfield><controlfield tag="008">230226s2021 xx \|\|\|\|\|o 00\| \|\|eng c</controlfield><datafield tag="024" ind1="7" ind2=" "><subfield code="a">10.3390/fi13120300</subfield><subfield code="2">doi</subfield></datafield><datafield tag="035" ind1=" " ind2=" "><subfield code="a">(DE-627)DOAJ018814255</subfield></datafield><datafield tag="035" ind1=" " ind2=" "><subfield code="a">(DE-599)DOAJeefa9c16defa469bb5e7866b8829d3b6</subfield></datafield><datafield tag="040" ind1=" " ind2=" "><subfield code="a">DE-627</subfield><subfield code="b">ger</subfield><subfield code="c">DE-627</subfield><subfield code="e">rakwb</subfield></datafield><datafield tag="041" ind1=" " ind2=" "><subfield code="a">eng</subfield></datafield><datafield tag="050" ind1=" " ind2="0"><subfield code="a">T58.5-58.64</subfield></datafield><datafield tag="100" ind1="0" ind2=" "><subfield code="a">Junhyung Kwon</subfield><subfield code="e">verfasserin</subfield><subfield code="4">aut</subfield></datafield><datafield tag="245" ind1="1" ind2="0"><subfield code="a">Improving the Robustness of Model Compression by On-Manifold Adversarial Training</subfield></datafield><datafield tag="264" ind1=" " ind2="1"><subfield code="c">2021</subfield></datafield><datafield tag="336" ind1=" " ind2=" "><subfield code="a">Text</subfield><subfield code="b">txt</subfield><subfield code="2">rdacontent</subfield></datafield><datafield tag="337" ind1=" " ind2=" "><subfield code="a">Computermedien</subfield><subfield code="b">c</subfield><subfield code="2">rdamedia</subfield></datafield><datafield tag="338" ind1=" " ind2=" "><subfield code="a">Online-Ressource</subfield><subfield code="b">cr</subfield><subfield code="2">rdacarrier</subfield></datafield><datafield tag="520" ind1=" " ind2=" "><subfield code="a">Despite the advance in deep learning technology, assuring the robustness of deep neural networks (DNNs) is challenging and necessary in safety-critical environments, including automobiles, IoT devices in smart factories, and medical devices, to name a few. Furthermore, recent developments allow us to compress DNNs to reduce the size and computational requirements of DNNs to fit them into small embedded devices. However, how robust a compressed DNN can be has not been well studied in addressing its relationship to other critical factors, such as prediction performance and model sizes. In particular, existing studies on robust model compression have been focused on the robustness against off-manifold adversarial perturbation, which does not explain how a DNN will behave against perturbations that follow the same probability distribution as the training data. This aspect is relevant for on-device AI models, which are more likely to experience perturbations due to noise from the regular data observation environment compared with off-manifold perturbations provided by an external attacker. Therefore, this paper investigates the robustness of compressed deep neural networks, focusing on the relationship between the model sizes and the prediction performance on noisy perturbations. Our experiment shows that on-manifold adversarial training can be effective in building robust classifiers, especially when the model compression rate is high.</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">model compression</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">adversarial robustness</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">robust compression</subfield></datafield><datafield tag="650" ind1=" " ind2="4"><subfield code="a">on-manifold perturbation</subfield></datafield><datafield tag="653" ind1=" " ind2="0"><subfield code="a">Information technology</subfield></datafield><datafield tag="700" ind1="0" ind2=" "><subfield code="a">Sangkyun Lee</subfield><subfield code="e">verfasserin</subfield><subfield code="4">aut</subfield></datafield><datafield tag="773" ind1="0" ind2="8"><subfield code="i">In</subfield><subfield code="t">Future Internet</subfield><subfield code="d">MDPI AG, 2010</subfield><subfield code="g">13(2021), 12, p 300</subfield><subfield code="w">(DE-627)610604147</subfield><subfield code="w">(DE-600)2518385-0</subfield><subfield code="x">19995903</subfield><subfield code="7">nnns</subfield></datafield><datafield tag="773" ind1="1" ind2="8"><subfield code="g">volume:13</subfield><subfield code="g">year:2021</subfield><subfield code="g">number:12, p 300</subfield></datafield><datafield tag="856" ind1="4" ind2="0"><subfield code="u">https://doi.org/10.3390/fi13120300</subfield><subfield code="z">kostenfrei</subfield></datafield><datafield tag="856" ind1="4" ind2="0"><subfield code="u">https://doaj.org/article/eefa9c16defa469bb5e7866b8829d3b6</subfield><subfield code="z">kostenfrei</subfield></datafield><datafield tag="856" ind1="4" ind2="0"><subfield code="u">https://www.mdpi.com/1999-5903/13/12/300</subfield><subfield code="z">kostenfrei</subfield></datafield><datafield tag="856" ind1="4" ind2="2"><subfield code="u">https://doaj.org/toc/1999-5903</subfield><subfield code="y">Journal toc</subfield><subfield code="z">kostenfrei</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_USEFLAG_A</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">SYSFLAG_A</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_DOAJ</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_11</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_20</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_22</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_23</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_24</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_39</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_40</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_60</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_62</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_63</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_65</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_69</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_70</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_73</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_95</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_105</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_110</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_151</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_161</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_170</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_213</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_230</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_285</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_293</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_370</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_602</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_2009</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_2014</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_2111</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_2129</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4012</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4037</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4046</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4112</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4125</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4126</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4249</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4305</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4306</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4307</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4313</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4322</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4323</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4324</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4325</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4326</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4335</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4338</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4367</subfield></datafield><datafield tag="912" ind1=" " ind2=" "><subfield code="a">GBV_ILN_4700</subfield></datafield><datafield tag="951" ind1=" " ind2=" "><subfield code="a">AR</subfield></datafield><datafield tag="952" ind1=" " ind2=" "><subfield code="d">13</subfield><subfield code="j">2021</subfield><subfield code="e">12, p 300</subfield></datafield></record></collection>
score	7.399809

Nicht das Richtige dabei?

Schreiben Sie uns!

Improving the Robustness of Model Compression by On-Manifold Adversarial Training

Nicht das Richtige dabei?

Zugang & Verfügbarkeit

Vorhandene Bände

Nicht das Richtige dabei?