Privacy-Preserving Phishing Web Page Classification Via Fully Homomorphic Encryption

Chou, Edward; Gururajan, Arun; Laine, Kim; Goel, Neerja; Bertiger, Anna; Stokes, Jack W.

doi:10.1109/icassp40776.2020.9053729

Cited by 15 publications

(5 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In recent years, the research increasingly focused on applying privacy-preserving cryptographic protocols and primitives to combat privacy issues during the training and classification of machine learning models. More specifically, secure multiparty computation (MPC) has been applied to both training [21][22][23] and classification [24][25][26][27], whereas solutions including HE have mostly only successfully been applied to classification tasks [28][29][30][31][32][33]. Training machine learning models on homomorphically encrypted inputs is still an ongoing research questions.…”

Section: Related Workmentioning

confidence: 99%

CryptoTL: Private, efficient and secure transfer learning

Walch¹,

Sousa²,

Helminger³

et al. 2022

Preprint

View full text Add to dashboard Cite

Big data has been a pervasive catchphrase in recent years, but dealing with data scarcity has become a crucial question for many real-world deep learning (DL) applications. A popular methodology to efficiently enable the training of DL models to perform tasks in scenarios where only a small dataset is available is transfer learning (TL). TL allows knowledge transfer from a general domain to a specific target one; however, such a knowledge transfer may put privacy at risk when it comes to sensitive or private data. With CryptoTL we introduce a solution to this problem, and show for the first time a cryptographic privacy-preserving TL approach based on homomorphic encryption that is efficient and feasible for real-world use cases. We demonstrate this by focusing on classification tasks with small datasets and show the applicability of our approach for sentiment analysis. Additionally we highlight how our approach can be combined with differential privacy to further increase the security guarantees. Our extensive benchmarks show that using CryptoTL leads to high accuracy while still having practical fine-tuning and classification runtimes despite using homomorphic encryption. Concretely, one forward-pass through the encrypted layers of our setup takes roughly 1 s on a notebook CPU.Preprint. Under review.

show abstract

Section: Related Workmentioning

confidence: 99%

CryptoTL: Private, efficient and secure transfer learning

Walch¹,

Sousa²,

Helminger³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…After the successful implementation of Craig Gentry's (Stanford Ph.D. thesis 2009) work, homomorphic operations have become an important, futuristic technique in cloud computing. Some of the recent works on homomorphic encryptions are Cominetti et al [35], Chou et al [36], and Turan et al [37]. [38] and Elgamal encryption [39] in which the plaintexts are encoded in the exponents.…”

Section: Preliminaries: Homomorphic Encryptionmentioning

confidence: 99%

FogDedupe: A Fog-Centric Deduplication Approach Using Multi-Key Homomorphic Encryption Technique

Yoosuf

Muralidharan

Shitharth³

et al. 2022

Journal of Sensors

View full text Add to dashboard Cite

The advancements in communication technologies and a rapid increase in the usage of IoT devices have resulted in an increased data generation rate. Storing, managing, and processing large quantities of unstructured data generated by IoT devices remain a huge challenge to cloud service providers (CSP). To reduce the storage overhead, CSPs implement deduplication algorithms on the cloud storage servers. It identifies and eliminates the redundant data blocks. However, implementing post-progress deduplication schemes does not address the bandwidth issues. Also, existing convergent key-based deduplication schemes are highly vulnerable to confirmation of file attacks (CFA) and can leak confidential information. To overcome these issues, FogDedupe, a fog-centric deduplication framework, is proposed. It performs source-level deduplication on the fog nodes to reduce the bandwidth usage and post-progress deduplication to improve the cloud storage efficiency. To perform source-level deduplication, a distributed index table is created and maintained in the fog nodes, and post-progress deduplication is performed using a multi-key homomorphic encryption technique. To evaluate the proposed FogDedupe framework, a testbed environment is created using the open-source Eucalyptus v.4.2.0 software and fog project v1.5.9 package. The proposed scheme tightens the security against CFA attacks and improves the storage overhead by 27% and reduces the deduplication latency by 12%.

show abstract

“…Moreover, the phishing URL database [2,21,24] that stores the observed phishing attacks provides an ideal testbed for the deep-learning-based URL classification task with a relatively closed environment. Various deep-learning methods such as CNN [5,6] and its modification [18,25,26] are proposed, as well as the LSTM-based generative adversarial network (GAN) [17] for exploiting the class imbalance issue by generating phishing URLs.…”

Section: Related Workmentioning

confidence: 99%

Deep Character-Level Anomaly Detection Based on a Convolutional Autoencoder for Zero-Day Phishing URL Detection

Cho

2021

Electronics

View full text Add to dashboard Cite

Considering the fatality of phishing attacks, the data-driven approach using massive URL observations has been verified, especially in the field of cyber security. On the other hand, the supervised learning approach relying on known attacks has limitations in terms of robustness against zero-day phishing attacks. Moreover, it is known that it is critical for the phishing detection task to fully exploit the sequential features from the URL characters. Taken together, to ensure both sustainability and intelligibility, we propose the combination of a convolution operation to model the character-level URL features and a deep convolutional autoencoder (CAE) to consider the nature of zero-day attacks. Extensive experiments on three real-world datasets consisting of 222,541 URLs showed the highest performance among the latest deep-learning methods. We demonstrated the superiority of the proposed method by receiver-operating characteristic (ROC) curve analysis in addition to 10-fold cross-validation and confirmed that the sensitivity improved by 3.98% compared to the latest deep model.

show abstract

Privacy-Preserving Phishing Web Page Classification Via Fully Homomorphic Encryption

Cited by 15 publications

References 12 publications

CryptoTL: Private, efficient and secure transfer learning

CryptoTL: Private, efficient and secure transfer learning

FogDedupe: A Fog-Centric Deduplication Approach Using Multi-Key Homomorphic Encryption Technique

Deep Character-Level Anomaly Detection Based on a Convolutional Autoencoder for Zero-Day Phishing URL Detection

Contact Info

Product

Resources

About