%PDF-1.4
%
1 0 obj
<>stream
application/pdfIEEEIEEE Access;2019;7; ;10.1109/ACCESS.2019.2951526Deep neural networksecurityadversarial attacksdefensesparse autoencoderdenoisingDDSA: A Defense Against Adversarial Attacks Using Deep Denoising Sparse AutoencoderYassine BakhtiSid Ahmed FezzaWassim HamidoucheOlivier Deforges
IEEE Access160397 2019710.1109/ACCESS.2019.2951526160407
endstream
endobj
2 0 obj
<>/Rect[132.959 110.697 156.531 121.597]/H/I>>
endobj
3 0 obj
[571 564 502 430 437 430 520 440 300 492 547 686 472 426 600 545 534 433 554 577 588 704 655 452 590 834 547 524 562 1043 1043 1043 1043 319 319 373 373 642 804 802 796 762 832 762 740 794 767 275 331 780 470 780]
endobj
4 0 obj
[/View/Design]
endobj
5 0 obj
<>/ProcSet[/PDF/Text]/ColorSpace<>/Font<>>>/MediaBox[0 0 576 782.929]/Annots[21 0 R 22 0 R 23 0 R 24 0 R 25 0 R 26 0 R 27 0 R]/Rotate 0>>
endobj
6 0 obj
<>
endobj
7 0 obj
<>stream
HWY+z:%:56$F,G-#>UQL#vEXONNxAq~W绷Wyl;?}'4r''|/YP:yʕ~᯼,Xfq]{}7ߋ=izcyg]5Mt*][0⹗}՜)r-
0 ݶqϤ/L8X$$+)[4h_͓Fi
w}viANZ~
iœw'=G!ltL{GaMc:=]{&
X_VX~cX_)p8P(RtURd>˯P
#+FmcdHw,`GM
Ҝ݅?ɄvAFw@Occ Vп\?U]
ݺfy|,ʈp>,~ӫCS_~IQ`%1bx`!sX#G`~0d?%QQB#JL>a]-l{|Ԧ<a#E bKضx{K
^h<8g[m
"?@wk_egvAB6cccE秱zIpIfBյbP_$8h__m9%)b@Ɔf\,9M]\()62F][%8HᢇYA-d{{W3XwC={J/ݦ
zp"S2BibO(,/8ںTRU2oЬ