Texception: A Character/Word-Level Deep Learning Model for Phishing URL Detection

Tajaddodianfar, Farid; Stokes, Jack W.; Gururajan, Arun

doi:10.1109/icassp40776.2020.9053670

Cited by 54 publications

(39 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The Monte Carlo Search-based triplet network is the latest implementation of deep metric learning [15] and it achieved the better performance compared to the base network with 0.9673 accuracy and 0.9227 recall. The texception network optimized to model the character and word-level features of phishing URLs achieved 0.9765 accuracy and 0.9462 recall [2], which are improved performance compared to the metric learning approach.…”

Section: Resultsmentioning

confidence: 95%

“…Deep learning-based approaches to detect phishing attacks are drawing attention from major societies in terms of data, structure, and learning strategy [9]. The improvement of character-level CNN studied in the field of sentiment classification was optimized for phishing URL detection by simultaneously modeling the word-level features [2]. In the case of modifying the data sampling algorithm, the phishing classification task was used as a benchmark task [10].…”

Section: Related Workmentioning

confidence: 99%

“…In the field of information security, an inductive approach to learn the features from phishing attacks based on a high-performance deep learning model with massive known-attack observations was introduced. Among the most prominent, the combination of a convolutional neural network (CNN) and a recurrent neural network has been found to significantly improve the detection performance by explicitly modeling the character and word features of phishing attacks [1], [2].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Integrating Deep Learning with First-Order Logic Programmed Constraints for Zero-Day Phishing Attack Detection

Cho

2021

ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

Considering the fatality of phishing attacks that are emphasized by many organizations, the inductive learning approach using reported malicious URLs has been verified in the field of deep learning. However, the deep learning-based method mainly focused on the fitting of a classification task via historical URL observation shows a limitation of recall due to the characteristics of zero-day attack. In order to model the nature of a zero-day phishing attack in which URL addresses are generated and discarded immediately, an approach that utilizes the expert knowledge is promising. We introduce the integration method of deep learning and logic programmed domain knowledge to inject the real-world constraints. We design neural and logic classifiers and propose the joint learning method of each component based on the traditional neuro-symbolic integration. Extensive experiments on three real-world datasets consisting of 222,541 URLs showed the highest recall among the latest deep learning methods, despite the hostile class-imbalanced condition. We demonstrate that the optimized weighting between neural and logic component has an effect of improving the recall over 3% compared to the existing methods.

show abstract

Section: Resultsmentioning

confidence: 95%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Integrating Deep Learning with First-Order Logic Programmed Constraints for Zero-Day Phishing Attack Detection

Cho

2021

ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

View full text Add to dashboard Cite

show abstract

“…The majority of the current research in deep learning-based phishing detection focuses mainly on optimizing the operation of the neural network [16]. In particular, the comparative study in [17] proves the superiority of the ensemble approach based on CNN variations.…”

Section: Related Workmentioning

confidence: 93%

Ensemble Learning of Deep URL Features based on Convolutional Neural Network for Phishing Attack Detection

Bu¹,

Kim²

2021

Adv. sci. technol. eng. syst. j.

View full text Add to dashboard Cite

The deep learning-based URL classification approach using massive observations has been verified especially in the field of phishing attack detection. Various improvements have been achieved through the modeling of character and word sequence of URL based on convolutional and recurrent neural networks, and it has been proven that an ensemble approach of each model has the best performance. However, existing ensemble methods have limitations in effectively fusing the nonlinear correlation between heterogeneous features extracted from characters and the sequence of sub-domains. In this paper, we propose a convolutional network-based ensemble learning approach to systematically fuse syntactic and semantic features for phishing URL detection. By learning the weights that integrating the heterogeneous features extracted from the URL, an ensemble rule that guarantees the best performance was obtained. A total of 45,000 benign URLs and 15,000 phishing URLs were collected and 10-fold cross-validation was conducted for quantitative validation. The obtained classification accuracy of 0.9804 indicates that the proposed method outperforms the existing machine learning algorithms and provides plausible solution for phishing URL detection. We demonstrated the superiority of the proposed method by receiver-operating characteristic (ROC) curve analysis and the case analysis and confirmed that the accuracy improved by 1.93% compared to the latest deep model.

show abstract

“…Hence, the comparison has been made with less suitable state-of-art methods. The comparison has been drawn with URLNet [88] , Texception [89] , Triple Network [90] , and Monte Carlo [91] . [91] is the most recent deep learning-based implementation.…”

Section: Performance Analysismentioning

confidence: 99%

Detecting Covid-19 chaos driven phishing/malicious URL attacks by a fuzzy logic and data mining based intelligence system

Zahra

Chishti²,

Baba³

et al. 2022

Egyptian Informatics Journal

View full text Add to dashboard Cite

With confusion and uncertainty ruling the world, 2020 created near-perfect conditions for cybercriminals. As businesses virtually eliminated in-person experiences, the COVID-19 pandemic changed the way we live and caused a mass migration to digital platforms. However, this shift also made people more vulnerable to cyber-crime. Victims are being targeted by attackers for their credentials or financial rewards, or both. This is because the Internet itself is inherently difficult to secure, and the attackers can code in a way that exploits its flaws. Once the attackers gain root access to the devices, they have complete control and can do whatever they want. Consequently, taking advantage of highly unprecedented circumstances created by the Covid-19 event, cybercriminals launched massive phishing, malware, identity theft, and ransomware attacks. Therefore, if we wish to save people from these frauds in times when millions have already been tipped into poverty and the rest are trying hard to sustain, it is imperative to curb these attacks and attackers. This paper analyses the impact of Covid-19 on various cyber-security related aspects and sketches out the timeline of Covid-19 themed cyber-attacks launched globally to identify the modus operandi of the attackers and the impact of attacks. It also offers a thoroughly researched set of mitigation strategies which can be employed to prevent the attacks in the first place. Moreover, this manuscript proposes a fuzzy logic and data mining-based intelligence system for detecting Covid-19 themed malicious URL/phishing attacks. The performance of the system has been evaluated against various malicious/phishing URLs, and it was observed that the proposed system is a viable solution to this problem.

show abstract

Texception: A Character/Word-Level Deep Learning Model for Phishing URL Detection

Cited by 54 publications

References 6 publications

Integrating Deep Learning with First-Order Logic Programmed Constraints for Zero-Day Phishing Attack Detection

Integrating Deep Learning with First-Order Logic Programmed Constraints for Zero-Day Phishing Attack Detection

Ensemble Learning of Deep URL Features based on Convolutional Neural Network for Phishing Attack Detection

Detecting Covid-19 chaos driven phishing/malicious URL attacks by a fuzzy logic and data mining based intelligence system

Contact Info

Product

Resources

About