117 Hits in 4.2 sec

An End-to-End Attack on Text-based CAPTCHAs Based on Cycle-Consistent Generative Adversarial Network [article]

Chunhui Li, Xingshu Chen, Haizhou Wang, Yu Zhang, Peiming Wang
2020 arXiv   pre-print
In this paper, we propose an efficient and simple end-to-end attack method based on cycle-consistent generative adversarial networks.  ...  Firstly, we train CAPTCHA synthesizers based on the cycle-GAN to generate some fake samples. Basic recognizers based on the convolutional recurrent neural network are trained with the fake data.  ...  The authors thank anonymous reviewers for their helpful comments to improve the paper.  ... 
arXiv:2008.11603v1 fatcat:5b5kw3in25dlpg44nphwa2ghyu

Securing IoT Devices: A Robust and Efficient Deep Learning with a Mixed Batch Adversarial Generation Process for CAPTCHA Security Verification

Stephen Dankwa, Lu Yang
2021 Electronics  
The goal of this current study is to perform security vulnerability verification on adversarial text-based CAPTCHA, based on attacker–defender scenarios.  ...  An adversarial machine learning attack is a good practice to adopt, to strengthen the security of text-based CAPTCHA (Completely Automated Public Turing test to tell Computers and Humans Apart), to withstand  ...  The authors express their sincere thanks to Gao Bin for his advice. Lastly, the authors thank Eric Affum, Lisa Agyemang, Edward Donkor, and Vida Sarpong for their encouragement.  ... 
doi:10.3390/electronics10151798 fatcat:e4khz6abfvglteflhgnvzwrbiy

CAPTCHA Image Generation Systems Using Generative Adversarial Networks

Hyun KWON, Yongchul KIM, Hyunsoo YOON, Daeseon CHOI
2018 IEICE transactions on information and systems  
We propose new CAPTCHA image generation systems by using generative adversarial network (GAN) techniques to strengthen against CAPTCHA solvers.  ...  The results show that the proposed schemes improve the resistance to the CAPTCHA solver by over 67.1% and 89.8% depending on the system. key words: CAPTCHA, generative adversarial network, deep convolutional  ...  Generative adversarial networks (GANs) have been recently introduced as a generative model for machine learning [7] .  ... 
doi:10.1587/transinf.2017edl8175 fatcat:o3xjjrncjzfqlpkmjwlftogvii

AniCAP: An Animated 3D CAPTCHA Scheme Based on Motion Parallax [chapter]

Yang-Wai Chow, Willy Susilo
2011 Lecture Notes in Computer Science  
To date, numerous CAPTCHA schemes have been proposed and deployed on various websites to secure online services from abuse by automated programs.  ...  However, many of these CAPTCHAs have been found to suffer from design flaws that can be exploited to break the CAPTCHA.  ...  While there are three main categories of CAPTCHAs; namely, text-based CAPTCHAs, audio-based CAPTCHAs and image-based CAPTCHAs, this paper focuses on text-based CAPTCHAs.  ... 
doi:10.1007/978-3-642-25513-7_18 fatcat:y275673i5vdepelgym34b7mwla

I am Robot: (Deep) Learning to Break Semantic Image CAPTCHAs

Suphannee Sivakorn, Iasonas Polakis, Angelos D. Keromytis
2016 2016 IEEE European Symposium on Security and Privacy (EuroS&P)  
Recent work, however, presented a generic attack that can be applied to any text-based captcha scheme. Fittingly, Google recently unveiled the latest version of reCaptcha.  ...  Based on our experimental findings, we propose a series of safeguards and modifications for impacting the scalability and accuracy of our attacks.  ...  Acknowledgements We would like to thank the anonymous reviewers, as well as Fabian Monrose and Michalis Polychronakis for their comments on previous drafts of this paper.  ... 
doi:10.1109/eurosp.2016.37 dblp:conf/eurosp/SivakornPK16 fatcat:u2gegsjtfnd3beflnxrem6idja

Designing Proof of Human-Work Puzzles for Cryptocurrency and Beyond [chapter]

Jeremiah Blocki, Hong-Sheng Zhou
2016 Lecture Notes in Computer Science  
By contrast, CAPTCHAs are only difficult for other computers to solve -not for the computer that generated the puzzle.  ...  of the Fourteenth IACR Theory of Cryptography Conference (TCC 2016-B). 1 Technically byzantine agreement is only possible when the adversary has less than 50% of the hashing power and the network has  ...  Acknowledgments: The authors thank paper shepherd Peter Gaži for his very constructive feedback which helped us to improve the quality of the paper.  ... 
doi:10.1007/978-3-662-53644-5_20 fatcat:m4wa3kkx6naglmtriqqbzjypk4

CACTI: Captcha Avoidance via Client-side TEE Integration [article]

Yoshimichi Nakatsuka, Ercan Ozturk, Andrew Paverd, Gene Tsudik
2020 arXiv   pre-print
These rate-proofs can be sent to web servers in lieu of solving CAPTCHAs.  ...  Using client-side TEEs, CACTI allows legitimate clients to generate unforgeable rate-proofs demonstrating how frequently they have performed specific actions.  ...  text recognition CAPTCHAs, as well as challenge-based CAPTCHAs consisting of simple algebraic operations on random integers.  ... 
arXiv:2007.10397v1 fatcat:f2267spyzjgovjws5ews7ccmey

Machine Learning Use Cases in Cybersecurity

S.М. Avdoshin, A.В. Lazarenko, N.I. Chichileva, P.А. Naumov, P.G. Klyucharev
2019 Proceedings of the Institute for System Programming of RAS  
Moreover, such technologies can be used by intruders to attack computer systems.  ...  The goal of this paper to explore machine learning usage in cybersecurity and cyberattack and provide a model of machine learning-powered attack.  ...  MalGAN is a generative adversarial network-based algorithm that generates adversarial malware examples that are able to bypass black-box ML-based detection models.  ... 
doi:10.15514/ispras-2019-31(5)-15 fatcat:dfnfslb4hffvrlmb65gkvrtrsm

Protecting Web 2.0 Services from Botnet Exploitations

Nguyen H. Vo, Josef Pieprzyk
2010 2010 Second Cybercrime and Trustworthy Computing Workshop  
Initially the EZ-Gimpy Captcha worked well but by the end of 2003, the next generation of bots and malware was able to read the text.Figure 5.1 show some samples of early Captcha.Modern Text-based CaptchaTo  ...  , we can improve Captcha security by using image-based Captcha instead of text-based Captcha.  ... 
doi:10.1109/ctc.2010.10 fatcat:cfoopseg4vc3lf7sfd2twrmfoy

Deep Learning in Information Security [article]

Stefan Thaler, Vlado Menkovski, Milan Petkovic
2018 arXiv   pre-print
Based on an analysis of our reviewed papers, we point out shortcomings of DL-methods to those requirements and discuss further research opportunities.  ...  If DL-methods succeed to solve problems on a data type in one domain, they most likely will also succeed on similar data from another domain.  ...  Their system is able to pass the reCAPTCHA challenge in 61.2% of the attacks. Finally, Osadachy et al. propose a CAPTCHA generation scheme, that is robust to DL-based CAPTCHA attacks [104].  ... 
arXiv:1809.04332v1 fatcat:xfb7lgrkw5cirdl3qvmg3ssnbi

Bicycle Attacks Considered Harmful: Quantifying the Damage of Widespread Password Length Leakage [article]

Benjamin Harsha, Robert Morton, Jeremiah Blocki, John Springer, Melissa Dark
2020 arXiv   pre-print
In an observational study, we find that most of the Alexa top 100 rates sites are vulnerable to bicycle attacks meaning that an eavesdropping attacker can infer the exact length of a password based on  ...  (NDSS 2016) to protect user accounts. The LinkedIn frequency corpus is based on over 170 million passwords making it the largest frequency corpus publicly available to password researchers.  ...  GMail uses CAPTCHAs [71] to rate limit an online attacker i.e., by requiring the attacker to solve a CAPTCHA challenge 3 after one or more incorrect login attempts.  ... 
arXiv:2002.01513v1 fatcat:qoqragkzezdx7c75apdthvwd2q

Adversarial Machine Learning Attacks and Defense Methods in the Cyber Security Domain [article]

Ihai Rosenberg and Asaf Shabtai and Yuval Elovici and Lior Rokach
2021 arXiv   pre-print
This paper is the first to discuss the unique challenges of implementing end-to-end adversarial attacks in the cyber security domain, map them in a unified taxonomy, and use the taxonomy to highlight future  ...  First, the adversarial attack methods are characterized based on their stage of occurrence, and the attacker's goals and capabilities.  ...  Three different black-box attacks were attempted by the adversary: an attack based on zeroth-order optimization (ZOO; see Section 3), an attack based on a GAN (similar to the one proposed by [74] ), and  ... 
arXiv:2007.02407v3 fatcat:rj3qomvg4bfb5p3atsct4winji

Secure Distributed Human Computation [chapter]

Craig Gentry, Zulfikar Ramzan, Stuart Stubblebine
2009 Lecture Notes in Computer Science  
We consider the general paradigm of using large-scale distributed computation to solve difficult problems, but where humans can act as agents and provide candidate solutions.  ...  This paper introduces a line of research on secure distributed human computation.  ...  The distribution network consists of computers that primarily host problems and schedule these problems to clients.  ... 
doi:10.1007/978-3-642-04904-0_24 fatcat:4wpkaotzfvcdznl5cv2dkvqj2a

Spread Identity: A new dynamic address remapping mechanism for anonymity and DDoS defense

Dhananjay Phatak, Alan T. Sherman, Nikhil Joshi, Bhushan Sonawane, Vivek G. Relan, Amol Dawalbhakta
2013 Journal of Computer Security  
With multiple IP addresses for the same destination, filtering based on destination addresses can block attackers without necessarily blocking legitimate users.  ...  Likewise, the installed base of operating systems running individual hosts in the internal network, together with the end-user application suites they support, remain untouched; thereby illustrating that  ...  Acknowledgments We are grateful to Xiaowei Yang and Xin Liu for providing their implementations of filter-based and capability-based DDoS protection mechanisms, which we used in our ns-2 simulations.  ... 
doi:10.3233/jcs-2012-0463 fatcat:fdg2i2hetnbqni2g4dznzrvzni

Exploiting open functionality in SMS-capable cellular networks

William Enck, Patrick Traynor, Patrick McDaniel, Thomas La Porta
2005 Proceedings of the 12th ACM conference on Computer and communications security - CCS '05  
In addition to voice services, these networks deliver alphanumeric text messages to the vast majority of wireless subscribers.  ...  We then characterize network behavior and explore a number of reconnaissance techniques aimed at effectively targeting attacks on these systems.  ...  comments on this paper.  ... 
doi:10.1145/1102120.1102171 dblp:conf/ccs/EnckTMP05 fatcat:5b2wefzjqrgo5pagr3w2xgaiwq
« Previous Showing results 1 — 15 out of 117 results