Building Robust Phishing Detection Systems with Deep Learning

By: Jampula Navaneeth1

1Vel Tech University, Chennai, India

2International Center for AI and Cyber Security Research and Innovations, Asia University, Taiwan, Email: navaneethjampula@gmail.com

Abstract

As we can see now a days that protecting our personal data is more important, since all our transactions are going through online. The phishers main aim is to get sensitive data from users and then to blackmail them or to get bank details without knowing the users. To detect these types of frauds we can build a excellent detection system using deep learning. In this article we are going to see how deep learning will help in building a robust system and highlighting the challenges.

Keywords: Deep Learning, Robust system, Challenges

Introduction

Phishing is a form of fraud to obtain important information thus results to violation of privacy. The untrustworthy people, known as scammers or cyber-criminals, employ different methods for attacking the weak points of the network and people’s psychological dilemmas. The most frequent type of attack method is known as the social engineering which is a psychological manipulation of human beings into performing activities that put into peril personal credentials [1].Later, there is a type of machine learning known as deep learning employing artificial neural networks (ANNs). Neural network can also learn from large data sets experiencing will fullness in text and image classifications. The major objective of this research is to determine the best method of identifying email phishing using deep learning strategies [2].

Background

Phishing continues to be a powerful method in cybercrime since most victims cannot distinguish between a phishing website or email address. Some of the most rigorous hurdles that keep tripping up the attempts to counter cyber threats especially the phishing attack are lack of effective anti cyber threats solutions. AI is gradually being considered as the next frontier for reinforcing cyber security mechanisms [3].

Use of Deep Learning for Phishing Detection

In phishing detection feature deep learning was used to classify the email as a legitimate or a phishing email depending on the content of the message including the subject. Despite the advancement in the power of the computers, AI is an emerging field in CS since it is an effort in an attempt to mimic the intelligence of the computers with those of human beings. If anything, learning may be one of the things that define the human species out of all other living creatures [4].

It includes operations through deep artificial neural networks or deep reinforcement learning known as deep learning. In literal sense, the term ‘deep’ is used to express a parameter that is the number of levels that are built in the system. The methods applied in the deep learning are: Convection neural network, Recurrent neural network, Long short term memory network and Auto encoder. There are conditions that define the selection of the method, and they include: the use of the method and the level of data participation, there are conditions that define the selection of the method, and they include [4].

Deep Learning Approaches for Phishing Detection

There are various approaches that are used in the phishing detection process, but some of them are key approaches in the deep learning. In this article let’s see those key approaches for phishing detection.

  • Artificial Neural Networks (ANN) has the peculiar characteristic that it does not incorporate intermediate cycle connectors between its components. In this approach, the architecture primarily include fully connected layers only [4]. Regarding the parameters, the regulations are established with respect to the identified output of the loss function, while for each of the nodes, the direct linear computation is applied.
  • Convolutional Neural Networks (CNN) is yet another variety of deep learning algorithms that can operate with very little pre-processing. They use a multilayer perceptron kind and perform convolution operations by visualizing these as window moving operations on a matrix [2].
  • The second subcategory of the deep learning approach is a Recurrent Neural Network [RON, it implies connections between the nodes forming a directed graph through a sequence. By that, it is quite easy to imagine and even easily illustrate RNN as several connected neural network blocks which are put in form of a chain [2], [4].
Figure 1: Phishing detection testing model

Challenges

As an example, Convolutional Neural Networks (CNNs) are very useful for two-dimensional data that has gridlike structures such as images and videos because they are able to learn representations of image features in a hierarchical fashion. Conversely, Recurrent Neural Networks (RNNs) are best suited for sequential data and natural language processing [5].

While supervised deep learning has attracted considerable attention, it has the disadvantage of relying on large annotated datasets that are costly to obtain and time taking to curate [6]. Consequently, picking out the most fitting algorithm for a specific application in cybersecurity is a tricky affair. Wrong choice of algorithm will lead to uncertain results thereby wasting labour and compromising the effectiveness as well as accuracy of the model itself [3].

Conclusion

With this article I want to conclude that phishing attacks can be detected by using deep learning detection system. Building a robust phishing detection system with deep learning is a challenging work. By leveraging advanced CNN, ANN and RNN techniques, organizations will help in protecting from the phishing attackers. As technology increases deep learning will also grow with advanced techniques for future implementations.

References

  1. I. Saha, D. Sarma, R. J. Chakma, M. N. Alam, A. Sultana, and S. Hossain, “Phishing Attacks Detection using Deep Learning Approach,” in 2020 Third International Conference on Smart Systems and Inventive Technology (ICSSIT), Aug. 2020, pp. 1180–1185. doi: 10.1109/ICSSIT48917.2020.9214132.
  2. S. Atawneh and H. Aljehani, “Phishing Email Detection Model Using Deep Learning,” Electronics, vol. 12, no. 20, Art. no. 20, Jan. 2023, doi: 10.3390/electronics12204261.
  3. N. Q. Do, A. Selamat, O. Krejcar, E. Herrera-Viedma, and H. Fujita, “Deep Learning for Phishing Detection: Taxonomy, Current Challenges and Future Directions,” IEEE Access, vol. 10, pp. 36429–36463, 2022, doi: 10.1109/ACCESS.2022.3151903.
  4. O. K. Sahingoz, E. BUBEr, and E. Kugu, “DEPHIDES: Deep Learning Based Phishing Detection System,” IEEE Access, vol. 12, pp. 8052–8070, 2024, doi: 10.1109/ACCESS.2024.3352629.
  5. A. M. Widodo et al., “Port-to-Port Expedition Security Monitoring System Based on a Geographic Information System,” IJDSGBT, vol. 13, no. 1, pp. 1–20, Jan. 2024, doi: 10.4018/IJDSGBT.335897.
  6. M. Rahaman, B. Chappu, N. Anwar, and P. K. Hadi, “Analysis of Attacks on Private Cloud Computing Services that Implicate Denial of Services (DoS),” vol. 4, 2022.
  7. P. Chaudhary, B. B. Gupta, K. T. Chui and S. Yamaguchi, “Shielding Smart Home IoT Devices against Adverse Effects of XSS using AI model,” 2021 IEEE International Conference on Consumer Electronics (ICCE), Las Vegas, NV, USA, 2021, pp. 1-5, doi: 10.1109/ICCE50685.2021.9427591.
  8. Li, K. C., Gupta, B. B., & Agrawal, D. P. (Eds.). (2020). Recent advances in security, privacy, and trust for internet of things (IoT) and cyber-physical systems (CPS).

Cite As

Navaneeth J. (2024) Building Robust Phishing Detection Systems with Deep Learning, Insights2Techinfo, pp.1

78470cookie-checkBuilding Robust Phishing Detection Systems with Deep Learning
Share this:

Leave a Reply

Your email address will not be published.