Efficient Cross-Project Software Defect Prediction Based on Federated Meta-Learning

Haisong Chen,Linlin Yang,Aili Wang

DOI: https://doi.org/10.3390/electronics13061105

IF: 2.9

2024-03-18

Electronics

Abstract:Software defect prediction is an important part of software development, which aims to use existing historical data to predict future software defects. Focusing on the model performance and communication efficiency of cross-project software defect prediction, this paper proposes an efficient communication-based federated meta-learning (ECFML) algorithm. The lightweight MobileViT network is used as the meta-learner of the Model Agnostic Meta-Learning (MAML) algorithm. By learning common knowledge on the local data of multiple clients, and then fine-tuning the model, the number of unnecessary iterations is reduced, and communication efficiency is improved while reducing the number of parameters. The gradient information model is encrypted using the differential privacy of the Laplace mechanism, and the optimal privacy budget is determined through experiments. Experiments on three public datasets (AEEEM, NASA, and Relink) verified the effectiveness of ECFML in terms of parameter quantity, convergence, and model performance of cross-project software defect prediction.

engineering, electrical & electronic,computer science, information systems,physics, applied

What problem does this paper attempt to address?

This paper is primarily dedicated to addressing the issue of cross-project prediction in software defect prediction and proposes an efficient method based on Federated Meta-Learning (FML) to tackle this problem. Specifically, the paper aims to solve the following key issues:

1. **Improving Model Performance and Communication Efficiency**: Existing cross-project software defect prediction methods often face the problem of data distribution heterogeneity, meaning there is a significant difference in data distribution between the source project and the target project. This leads to poor prediction performance of models trained on the source project when applied to the target project. Additionally, existing methods suffer from large parameter sizes and low communication efficiency.

2. **Reducing Unnecessary Iterations**: By utilizing the lightweight MobileViT network as a meta-learner, it is possible to learn general knowledge on local data from multiple clients and reduce unnecessary iterations through model fine-tuning, thereby improving communication efficiency and reducing the number of parameters.

3. **Protecting Privacy and Ensuring Security**: To ensure the privacy and security of participants' data, the paper employs differential privacy technology under the Laplace mechanism to encrypt gradient information and determines the optimal privacy budget through experiments.

In summary, this research aims to propose an efficient cross-project software defect prediction algorithm based on Federated Meta-Learning (ECFML) to overcome the issues present in existing methods, achieving better model performance, higher communication efficiency, and stronger data privacy protection. The algorithm effectively addresses the aforementioned challenges by introducing the lightweight MobileViT network, reducing unnecessary iterations, and adopting differential privacy technology.