izpis_h1_title_alt

An experimental evaluation of adversarial examples and methods of defense
ID Šircelj, Jaka (Author), ID Skočaj, Danijel (Mentor) More about this mentor... This link opens in a new window

.pdfPDF - Presentation file, Download (2,67 MB)
MD5: FC97396BA6FEBA68500FE62024F04B99

Abstract
In this thesis we perform an experimental analysis and evaluation of different methods for creating adversarial examples, and learn how these affect different types of image classifiers, with the intent to obtain a better understanding of adversarial examples. The adversarial methods are hard to compare, since they use different types of parameters. We introduce a novel visualization technique, called accuracy-perturbation curve, that allows us to perform our comparison much more in depth, without the need to find optimal parameters. With this technique we also evaluate the successfulness of adversarial training as a defensive method. The results showed that radial basis function network classifiers possess an intrinsic property that makes them stronger on adversarial examples, compared to other classifiers, like CNNs, even though they perform poorly on clean images. Also, we noticed a weak correlation between the classifiers ability to generalize and its robustness against attacks.

Language:English
Keywords:adversarial examples, neural networks, deep learning, image classification
Work type:Master's thesis/paper
Organization:FRI - Faculty of Computer and Information Science
Year:2019
PID:20.500.12556/RUL-111417 This link opens in a new window
COBISS.SI-ID:1538387139 This link opens in a new window
Publication date in RUL:30.09.2019
Views:1745
Downloads:356
Metadata:XML DC-XML DC-RDF
:
Copy citation
Share:Bookmark and Share

Secondary language

Language:Slovenian
Title:Eksperimentalno ovrednotenje nasprotniških primerov in načinov obrambe
Abstract:
V tem delu opravimo eksperimentalno analizo in evalvacijo različnih metod generiranja nasprotniških primerov oz. evalviramo njihove vplive na različne tipe klasifikatorjev slik. Namen analize je bil pridobiti čim več znanja o nasprotniških primerih. Metode ustvarjanja nasprotniških primerov je zahtevno primerjati, ker vse uporabljajo drugačne tipe parametrov. Da se znebimo skrbi glede določanja optimalnih parametrov, uvedemo točnostno-perturbacijsko krivuljo, s katero lahko veliko bolj natančno ocenimo, koliko je klasifikator robusten pri obrambi oz. koliko je generator nasprotniških primerov uspešen pri napadu. S to krivuljo smo analizirali tudi obrambno metodo učenja na nasprotniških primerih. Rezultati kažejo, da so mreže z radialnimi baznimi funkcijami naravno bolj robustne proti takšnim napadom, tudi če v večini primerov niso primerne za klasifikacijo slik. Opazili smo še šibko korelacijo med zmožnostjo generalizacije klasifikatorjev ter njihovo odpornostjo pred nasprotniškimi primeri.

Keywords:nasprotniški primeri, nevronske mreže, globoko učenje, klasifikacija slik

Similar documents

Similar works from RUL:
Similar works from other Slovenian collections:

Back