Automated Bug Detection and Classification Using Deep Learning Techniques

Abstract

Automated bug detection and classification is a critical aspect of software development, as it enables early identification of issues, improving the quality and efficiency of software systems. In recent years, deep learning techniques have emerged as a promising approach to enhance the accuracy and efficiency of bug detection and classification. This paper explores the application of deep learning techniques for automating the bug detection and classification process, offering an innovative solution to traditional methods that rely on rule-based systems or manual inspection.

Deep learning models, particularly Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), and Transformer-based models, are discussed for their capabilities to analyze code patterns, identify bugs, and classify their severity. The paper also compares various deep learning architectures in terms of their performance and ability to handle real-world, large-scale datasets. The effectiveness of these models is evaluated through experiments using several benchmark datasets, demonstrating how they outperform traditional methods by offering faster, more reliable, and scalable solutions.

Furthermore, the paper discusses the importance of feature extraction, data preprocessing, and the challenges associated with training deep learning models for bug detection. It highlights the need for diverse datasets to account for the variety of bugs that can occur in different software applications. The results show that deep learning models, particularly those trained on code-level features, provide a promising alternative to manual bug classification and detection, significantly reducing the time and effort required for debugging.

This research highlights the potential for deep learning-based automated bug detection and classification systems to revolutionize software development practices, offering benefits such as improved code quality, faster debugging cycles, and more reliable software products.

DOI - https://doi.org/10.65525/SVUP.9788199778009.2026.77-87