MLPhishChain: a machine learning-based blockchain framework for reducing phishing threats DOI Creative Commons
Fouad Trad,

Elie Semaan-Nasr,

Ali Chehab

et al.

Frontiers in Blockchain, Journal Year: 2024, Volume and Issue: 7

Published: Dec. 12, 2024

Introduction Phishing attacks pose a significant threat to online security by deceiving users into divulging sensitive information through fraudulent websites. Traditional anti-phishing approaches are centralized and reactive, exhibiting critical limitations such as delayed detection, poor adaptability evolving threats, susceptibility data tampering, lack of transparency. Methods This paper presents MLPhishChain, decentralized application (DApp) that integrates blockchain technology with machine learning (ML) provide proactive transparent solution for URL verification. Users can submit URLs automated phishing analysis via an ML model, each URL’s status securely recorded on immutable ledger. To address the dynamic nature MLPhishChain features re-evaluation mechanism, enabling request updated assessments website content evolve. Additionally, system incorporates from external services (e.g., VirusTotal) offer multi-source validation risk, enhancing user confidence decision-making. Results The was built using Ganache Truffle, performance metrics were computed evaluate its efficacy in terms latency, scalability, resource consumption. indicate proposed achieves rapid verification low scales effectively handle increasing submissions, optimizes usage. Discussion By leveraging strengths intelligent algorithms, addresses shortcomings traditional methods. It delivers reliable adaptable capable addressing threats. approach establishes new standard characterized enhanced transparency, resilience, adaptability.

Language: Английский

Enhancing Privacy Policy Comprehension Through Privacify: A User-Centric Approach Using Advanced Language Models DOI

Justin Woodring,

K. Blake Perez, Aisha Ali-Gombe

et al.

Published: Jan. 1, 2024

Language: Английский

Citations

0
Investigating translation for Indic languages with BLOOMZ-3b through prompting and LoRA fine-tuning DOI Creative Commons

Aarathi Rajagopalan Nair,

Deepa Gupta,

B. Premjith

et al.

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: Oct. 15, 2024

In the domain of natural language processing, rise Large Language Models and Generative AI represents a noteworthy transition, enabling machines to understand generate text resembling that produced by humans. This research conducts thorough examination this transformative technology, with focus on its influence machine translation. The study explores translation landscape between English Indic languages, which include Hindi, Kannada, Malayalam, Tamil, Telugu. To address this, Model, BLOOMZ-3b, is utilized, has been primarily developed for generation task. Multiple prompting engineering techniques are prominently explored. further traverse fine-tuning BLOOMZ-3b model using Parameter Efficient Fine-Tuning technique called Low Rank Adaptation, aiming reduce computational complexity. Hence, combining innovative approaches model, it contributes continuous development technologies beyond traditional borders what can be done respect processing. regard, not only does shed light intricacy problems but also sets precedence optimizing or adapting big models various languages end up advancing Artificial Intelligence Natural Processing at large.

Language: Английский

Citations

0
Heuristic machine learning approaches for identifying phishing threats across web and email platforms DOI Creative Commons

Ramprasath Jayaprakash,

Krishnaraj Natarajan, Alfred Daniel

et al.

Frontiers in Artificial Intelligence, Journal Year: 2024, Volume and Issue: 7

Published: Oct. 21, 2024

Life has become more comfortable in the era of advanced technology this cutthroat competitive world. However, there are also emerging harmful technologies that pose a threat. Without doubt, phishing is one rising concerns leads to stealing vital information such as passwords, security codes, and personal data from any target node through communication hijacking techniques. In addition, attacks include delivering false messages originate trusted source. Moreover, attack aims get victim run malicious programs reveal confidential data, bank credentials, one-time user login credentials. The sole intention collect program-based attempts embedded URLs, emails, website-based attempts. Notably, proposed technique detects URL, email, attacks, which will be beneficial secure us scam Subsequently, pre-processed identify using cleaning, attribute selection, detected machine learning Furthermore, techniques use heuristic-based attacks. Admittedly, 56 features used analyze URL findings, experimental results show better accuracy 97.2%. Above all, for email detection obtain higher 97.4%. website 98.1%, 48 analysis.

Language: Английский

Citations

0
“Is this Site Legit?”: LLMs for Scam Website Detection DOI

Yuan-Chen Chang,

Esma Aı̈meur

Lecture notes in computer science, Journal Year: 2024, Volume and Issue: unknown, P. 230 - 245

Published: Nov. 26, 2024

Language: Английский

Citations

0
MLPhishChain: a machine learning-based blockchain framework for reducing phishing threats DOI Creative Commons
Fouad Trad,

Elie Semaan-Nasr,

Ali Chehab

et al.

Frontiers in Blockchain, Journal Year: 2024, Volume and Issue: 7

Published: Dec. 12, 2024

Introduction Phishing attacks pose a significant threat to online security by deceiving users into divulging sensitive information through fraudulent websites. Traditional anti-phishing approaches are centralized and reactive, exhibiting critical limitations such as delayed detection, poor adaptability evolving threats, susceptibility data tampering, lack of transparency. Methods This paper presents MLPhishChain, decentralized application (DApp) that integrates blockchain technology with machine learning (ML) provide proactive transparent solution for URL verification. Users can submit URLs automated phishing analysis via an ML model, each URL’s status securely recorded on immutable ledger. To address the dynamic nature MLPhishChain features re-evaluation mechanism, enabling request updated assessments website content evolve. Additionally, system incorporates from external services (e.g., VirusTotal) offer multi-source validation risk, enhancing user confidence decision-making. Results The was built using Ganache Truffle, performance metrics were computed evaluate its efficacy in terms latency, scalability, resource consumption. indicate proposed achieves rapid verification low scales effectively handle increasing submissions, optimizes usage. Discussion By leveraging strengths intelligent algorithms, addresses shortcomings traditional methods. It delivers reliable adaptable capable addressing threats. approach establishes new standard characterized enhanced transparency, resilience, adaptability.

Language: Английский

Citations

0