Adversarial Attacks and Defenses for Social Network Text Processing Applications: Techniques, Challenges and Future Research Directions

Alsmadi, Izzat; Ahmad, Kashif; Nazzal, Mahmoud; Alam, Firoj; Al-Fuqaha, Ala; Khreishah, Abdallah; Algosaibi, Abdulelah

doi:10.48550/arxiv.2110.13980

Cited by 7 publications

(5 citation statements)

References 109 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Ensemble methods, combining multiple models to improve detection accuracy, have also been highlighted for their effectiveness in reducing false positives [20], [21]. Natural Language Processing (NLP) techniques have been leveraged to understand the semantic content of messages, offering insights into the subtle cues used by spammers [22], [23], [24], [25], [26]. Anomaly detection approaches have been proposed to identify spam by detecting deviations from typical user behavior or message content [27], [28], [29].…”

Section: A Review Of Spam Detection Techniquesmentioning

confidence: 99%

Comparative Analysis of ChatGPT-4 and Google Gemini for Spam Detection on the SpamAssassin Public Mail Corpus

Mardiansyah,

Surya

2024

Preprint

View full text Add to dashboard Cite

This study addresses the critical challenge of spam detection in the realm of cybersecurity, motivated by the escalating sophistication of spamming techniques and their significant implications for communication security. With the advent of advanced artificial intelligence (AI) models, this research compares the efficacy of two leading models, ChatGPT-4 by OpenAI and Google Gemini, in identifying spam within the widely recognized SpamAssassin public mail corpus. Through a meticulous methodology that includes preprocessing of the dataset, application of standardized evaluation metrics (accuracy, precision, recall, and F1-score), and a detailed performance analysis, this study unveils the distinct capabilities of each model in spam detection. ChatGPT-4 demonstrates a balanced performance with high precision and recall, making it suitable for general spam detection tasks. In contrast, Google Gemini excels in recall, highlighting its potential in scenarios where capturing the maximum number of spam emails is paramount, despite a slightly higher tendency to misclassify legitimate emails as spam. These findings contribute valuable insights into the comparative strengths and application contexts of ChatGPT-4 and Google Gemini, offering a nuanced understanding of their roles in enhancing spam detection mechanisms. The study underscores the significance of selecting AI models based on specific spam detection needs and sets a foundation for future research aimed at advancing AI-driven spam detection solutions.

show abstract

Section: A Review Of Spam Detection Techniquesmentioning

confidence: 99%

Comparative Analysis of ChatGPT-4 and Google Gemini for Spam Detection on the SpamAssassin Public Mail Corpus

Mardiansyah,

Surya

2024

Preprint

View full text Add to dashboard Cite

show abstract

“…Adversarial learning specifically applied to OSN have been studied in different ways. The first one focuses on text processing applications [2], which is what event detection algorithms are. The second one is more specific, it is about evading spam detection [9].…”

Section: Related Workmentioning

confidence: 99%

Malicious Human Behaviour in Information System Security: Contribution to a Threat Model for Event Detection Algorithms

Casanove

Sèdes

2023

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Among the issues our community has to fix, information system security is a global concern both for the security of data and algorithms. The security of algorithms is dependent of the reliability of the input data. This reliability is questioned, especially when the data is generated by humans (or bots operated by humans), like in online social networks. Event detection algorithms are an example of technology using this type of data, but the question of the security is not systematically considered in this literature. We propose in this paper a first contribution to a threat model to overcome this problem. This threat model is composed of a description of the subject we are modelling, assumptions made, potential threats and defense strategies.

show abstract

“…Related surveys and differences with this survey. Reviews of adversarial examples attacking deep learning models for text applications already exist in the literature, such as [10,[12][13][14]; however, to the best of our knowledge, no comprehensive review has collected and summarized the efforts in this research direction while specifically focusing on sentiment analysis. This study attempts to cover this gap via three central questions:…”

Section: Introductionmentioning

confidence: 99%

A Survey of Adversarial Attacks: An Open Issue for Deep Learning Sentiment Analysis Models

Vázquez-Hernández,

Morales-Rosales,

Algredo-Badillo

et al. 2024

Applied Sciences

View full text Add to dashboard Cite

In recent years, the use of deep learning models for deploying sentiment analysis systems has become a widespread topic due to their processing capacity and superior results on large volumes of information. However, after several years’ research, previous works have demonstrated that deep learning models are vulnerable to strategically modified inputs called adversarial examples. Adversarial examples are generated by performing perturbations on data input that are imperceptible to humans but that can fool deep learning models’ understanding of the inputs and lead to false predictions being generated. In this work, we collect, select, summarize, discuss, and comprehensively analyze research works to generate textual adversarial examples. There are already a number of reviews in the existing literature concerning attacks on deep learning models for text applications; in contrast to previous works, however, we review works mainly oriented to sentiment analysis tasks. Further, we cover the related information concerning generation of adversarial examples to make this work self-contained. Finally, we draw on the reviewed literature to discuss adversarial example design in the context of sentiment analysis tasks.

show abstract

Adversarial Attacks and Defenses for Social Network Text Processing Applications: Techniques, Challenges and Future Research Directions

Cited by 7 publications

References 109 publications

Comparative Analysis of ChatGPT-4 and Google Gemini for Spam Detection on the SpamAssassin Public Mail Corpus

Comparative Analysis of ChatGPT-4 and Google Gemini for Spam Detection on the SpamAssassin Public Mail Corpus

Malicious Human Behaviour in Information System Security: Contribution to a Threat Model for Event Detection Algorithms

A Survey of Adversarial Attacks: An Open Issue for Deep Learning Sentiment Analysis Models

Contact Info

Product

Resources

About