Adversarial Attacks - Versionsgeschichte

C.heck am 17. Dezember 2020 um 15:43 Uhr

2020-12-17T15:43:08Z

← Nächstältere Version		Version vom 17. Dezember 2020, 17:43 Uhr
Zeile 10:		Zeile 10:

	https://github.com/zentralwerkstatt/adversarial/blob/master/adversarial.ipynb		https://github.com/zentralwerkstatt/adversarial/blob/master/adversarial.ipynb

			https://christophm.github.io/interpretable-ml-book/adversarial.html

			https://b-ok.cc/book/5260920/fee7e3 < book: Strengthening Deep Neural Networks: Making AI Less Susceptible to Adversarial Trickery

C.heck am 17. Dezember 2020 um 15:40 Uhr

2020-12-17T15:40:20Z

@@ Zeile 1: / Zeile 1: @@
 Blog von Francis Hunger und Flupke mit schönen beispielen...: http://adversarial.io/blog/allgemein/
 ----
 https://towardsdatascience.com/perhaps-the-simplest-introduction-of-adversarial-examples-ever-c0839a759b8d

C.heck am 14. Juni 2020 um 12:30 Uhr

2020-06-14T12:30:16Z

← Nächstältere Version		Version vom 14. Juni 2020, 14:30 Uhr
Zeile 3:		Zeile 3:
	https://towardsdatascience.com/perhaps-the-simplest-introduction-of-adversarial-examples-ever-c0839a759b8d		https://towardsdatascience.com/perhaps-the-simplest-introduction-of-adversarial-examples-ever-c0839a759b8d
	---		---
			* A Complete List of All (arXiv) Adversarial Example Papers https://nicholas.carlini.com/writing/2019/all-adversarial-example-papers.html
	* Praxis-Beispiele: https://boingboing.net/tag/adversarial-examples		* Praxis-Beispiele: https://boingboing.net/tag/adversarial-examples
	* https://bdtechtalks.com/2018/12/27/deep-learning-adversarial-attacks-ai-malware/		* https://bdtechtalks.com/2018/12/27/deep-learning-adversarial-attacks-ai-malware/

C.heck am 30. Mai 2020 um 17:47 Uhr

2020-05-30T17:47:48Z

← Nächstältere Version		Version vom 30. Mai 2020, 19:47 Uhr
Zeile 1:		Zeile 1:
			Blog von Francis Hunger und Flupke mit schönen beispielen...: http://adversarial.io/blog/allgemein/
			----
	https://towardsdatascience.com/perhaps-the-simplest-introduction-of-adversarial-examples-ever-c0839a759b8d		https://towardsdatascience.com/perhaps-the-simplest-introduction-of-adversarial-examples-ever-c0839a759b8d
	---		---

C.heck: /* BLACK BOX / WHITE BOX ATTACKS */

2020-01-29T14:24:43Z

BLACK BOX / WHITE BOX ATTACKS

← Nächstältere Version		Version vom 29. Januar 2020, 16:24 Uhr
Zeile 122:		Zeile 122:
	* https://github.com/bethgelab		* https://github.com/bethgelab
	* https://github.com/tensorflow/cleverhans		* https://github.com/tensorflow/cleverhans

			----
			=Detection through [https://exmediawiki.khm.de/exmediawiki/index.php?title=XAI_/_Language_Models&#XAI XAI]=

			When [https://exmediawiki.khm.de/exmediawiki/index.php?title=XAI_/_Language_Models&#XAI Explainability] Meets Adversarial Learning: Detecting Adversarial Examples using SHAP Signatures: https://arxiv.org/pdf/1909.03418.pdf

			Using [https://exmediawiki.khm.de/exmediawiki/index.php?title=XAI_/_Language_Models&#XAI Explainabilty] to Detect Adversarial Attacks: https://openreview.net/pdf?id=B1xu6yStPH

	[[Category:Hacking]]		[[Category:Hacking]]

C.heck am 29. Januar 2020 um 14:17 Uhr

2020-01-29T14:17:49Z

← Nächstältere Version		Version vom 29. Januar 2020, 16:17 Uhr
Zeile 1:		Zeile 1:
			https://towardsdatascience.com/perhaps-the-simplest-introduction-of-adversarial-examples-ever-c0839a759b8d
			---
	* Praxis-Beispiele: https://boingboing.net/tag/adversarial-examples		* Praxis-Beispiele: https://boingboing.net/tag/adversarial-examples
	* https://bdtechtalks.com/2018/12/27/deep-learning-adversarial-attacks-ai-malware/		* https://bdtechtalks.com/2018/12/27/deep-learning-adversarial-attacks-ai-malware/

C.heck am 20. September 2019 um 10:21 Uhr

2019-09-20T10:21:48Z

← Nächstältere Version		Version vom 20. September 2019, 12:21 Uhr
Zeile 7:		Zeile 7:
	* https://en.wikipedia.org/wiki/Deep_learning#Cyberthreat		* https://en.wikipedia.org/wiki/Deep_learning#Cyberthreat
	[[File:Adv-attack.png\|800]]		[[File:Adv-attack.png\|800]]

			=notes=
			https://github.com/tensorflow/cleverhans
	=WHITE BOX ATTACKS=		=WHITE BOX ATTACKS=
	* https://cv-tricks.com/how-to/breaking-deep-learning-with-adversarial-examples-using-tensorflow/		* https://cv-tricks.com/how-to/breaking-deep-learning-with-adversarial-examples-using-tensorflow/
Zeile 18:		Zeile 21:
	===Fast Gradient Sign Method(FGSM)===		===Fast Gradient Sign Method(FGSM)===
	FGSM is a single step attack, ie.. the perturbation is added in a single step instead of adding it over a loop (Iterative attack).		FGSM is a single step attack, ie.. the perturbation is added in a single step instead of adding it over a loop (Iterative attack).
			* https://www.tensorflow.org/beta/tutorials/generative/adversarial_fgsm
			* https://arxiv.org/pdf/1412.6572.pdf
			** https://github.com/soumyac1999/FGSM-Keras

	===Basic Iterative Method===		===Basic Iterative Method===

C.heck: /* Another Attack Against Driverless Cars */

2019-08-22T17:35:13Z

Another Attack Against Driverless Cars

← Nächstältere Version		Version vom 22. August 2019, 19:35 Uhr
Zeile 73:		Zeile 73:
	<u>Schneier News:</u>		<u>Schneier News:</u>

	In this piece of research, attackers [https://arxiv.org/pdf/1906.09765.pdf successfully attack] a driverless car system -- Renault Captur's "Level 0" autopilot (Level 0 systems advise human drivers but do not directly operate cars) -- by following them with drones that project images of fake road signs in 100ms bursts. The time is too short for human perception, but long enough to fool the autopilot's sensors.		In this piece of research, attackers [https://arxiv.org/pdf/1906.09765.pdf successfully attack] a driverless car system -- Renault Captur's "Level 0" autopilot (Level 0 systems advise human drivers but do not directly operate cars) -- by following them with drones that project images of fake road signs in 100ms bursts. The time is too short for human perception, but long enough to fool the autopilot's sensors.

	Boing Boing [https://boingboing.net/2019/07/06/flickering-car-ghosts.html post].		Boing Boing [https://boingboing.net/2019/07/06/flickering-car-ghosts.html post].

C.heck: /* Another Attack Against Driverless Cars */

2019-08-22T17:35:02Z

Another Attack Against Driverless Cars

← Nächstältere Version		Version vom 22. August 2019, 19:35 Uhr
Zeile 71:		Zeile 71:
	===Another Attack Against Driverless Cars===		===Another Attack Against Driverless Cars===

	~~[2019.07.31]~~ In this piece of research, attackers [https://arxiv.org/pdf/1906.09765.pdf successfully attack] a driverless car system -- Renault Captur's "Level 0" autopilot (Level 0 systems advise human drivers but do not directly operate cars) -- by following them with drones that project images of fake road signs in 100ms bursts. The time is too short for human perception, but long enough to fool the autopilot's sensors.		<u>Schneier News:</u>

			In this piece of research, attackers [https://arxiv.org/pdf/1906.09765.pdf successfully attack] a driverless car system -- Renault Captur's "Level 0" autopilot (Level 0 systems advise human drivers but do not directly operate cars) -- by following them with drones that project images of fake road signs in 100ms bursts. The time is too short for human perception, but long enough to fool the autopilot's sensors.

	Boing Boing [https://boingboing.net/2019/07/06/flickering-car-ghosts.html post].		Boing Boing [https://boingboing.net/2019/07/06/flickering-car-ghosts.html post].

C.heck: /* on computer vision */

2019-08-22T17:34:32Z

on computer vision

← Nächstältere Version		Version vom 22. August 2019, 19:34 Uhr
Zeile 68:		Zeile 68:
	* https://boingboing.net/2019/03/31/mote-in-cars-eye.html		* https://boingboing.net/2019/03/31/mote-in-cars-eye.html
	** Paper vom Forschungsteam: https://keenlab.tencent.com/en/whitepapers/Experimental_Security_Research_of_Tesla_Autopilot.pdf		** Paper vom Forschungsteam: https://keenlab.tencent.com/en/whitepapers/Experimental_Security_Research_of_Tesla_Autopilot.pdf

			===Another Attack Against Driverless Cars===

			[2019.07.31] In this piece of research, attackers [https://arxiv.org/pdf/1906.09765.pdf successfully attack] a driverless car system -- Renault Captur's "Level 0" autopilot (Level 0 systems advise human drivers but do not directly operate cars) -- by following them with drones that project images of fake road signs in 100ms bursts. The time is too short for human perception, but long enough to fool the autopilot's sensors.

			Boing Boing [https://boingboing.net/2019/07/06/flickering-car-ghosts.html post].

	----		----

	==on voice (ASR)==		==on voice (ASR)==
	* https://www.the-ambient.com/features/weird-ways-echo-can-be-hacked-how-to-stop-it-231		* https://www.the-ambient.com/features/weird-ways-echo-can-be-hacked-how-to-stop-it-231