CooBa: Cross-project Bug Localization via Adversarial Transfer Learning

Abstract

<jats:p>Bug localization plays an important role in software quality control. Many supervised machine learning models have been developed based on historical bug-fix information. Despite being successful, these methods often require sufficient historical data (i.e., labels), which is not always available especially for newly developed software projects. In response, cross-project bug localization techniques have recently emerged whose key idea is to transferring knowledge from label-rich source project to locate bugs in the target project. However, a major limitation of these existing techniques lies in that they fail to capture the specificity of each individual project, and are thus prone to negative transfer. To address this issue, we propose an adversarial transfer learning bug localization approach, focusing on only transferring the common characteristics (i.e., public information) across projects. Specifically, our approach (CooBa) learns the indicative public information from cross-project bug reports through a shared encoder, and extracts the private information from code files by an individual feature extractor for each project. CooBa further incorporates adversarial learning mechanism to ensure that public information shared between multiple projects could be effectively extracted. Extensive experiments on four large-scale real-world data sets demonstrate that the proposed CooBa significantly outperforms the state of the art techniques.</jats:p>

Uncertainty-wise software anti-patterns detection: A possibilistic evolutionary machine learning approach

“…However, we have not predicted the labels of the target software projects [103,104]). Hence, equipping our Bi-ADIPOK with a transfer learner capable of transferring knowledge from the considered source projects to target ones from various domains will be interesting [105,106,107]. 9, which are among the most common addressed anti-patterns in the research topics of software maintenance [108], [109], [110], [78], [111], [112], [72]: information is inherited by every class.…”

An approach of method‐level bug localization

“…In recent years, some researchers have proposed learningbased bug localization approaches [38,40,[44][45][46][47][58][59][60][61]. Unlike the aforementioned algorithms that explicitly pre-define features (e.g., TF-IDF) and then use the features to directly calculate the correlation between a new bug report and each source file, these learning-based algorithms train a neural network model from historical data (i.e., pairs of fixed bug report and corresponding buggy source file) to automatically learn the relevant features, and then predict the correlation between a new bug report and each source file using the trained model.…”

Fast Changeset-based Bug Localization with BERT

“…For instance, TRANP-CNN [17] is a recent technique that combines cross-project transfer learning and convolutional neural networks to achieve state-of-the-art performance on file-level bug localization. CooBa improves on TRANP-CNN by combining a shared encoder to capture cross-project with per-project features and using adversarial training to ensure that the per-project information remains unaffected by noise [63]. Lam et al's technique, DNNLOC, combines a deep neural network with the VSM in order to be effective across different types of similarity [25].…”