Home / Current Issue / Paper 1719451
Intelligent Software Defect Prediction Using Multimodal Deep Learning Through the Integration of Source Code, Software Metrics and Historical Development Data
Subject area: Science,Engineering and Technology · Area of research: Deep Learning
DOI: 10.64388/IREV10I1-1719451
Abstract
Software defect prediction has historically relied on a single view of source code — either handcrafted metrics, token sequences, or structural representations in isolation. Multimodal deep learning addresses the fundamental limitation that no single code view captures the full richness of software artifacts: their syntax, structure, semantics, history, and natural-language context. This report provides a comprehensive survey of multimodal deep learning approaches for software defect prediction, covering the major modalities (lexical/semantic, structural/AST, control-and-data-flow, metric-based, and natural language) and the fusion architectures that combine them — concatenation, attention-gating, cross-attention, and contrastive multi-view learning. We synthesize findings from over 50 recent publications (2020-2026), including GMCA-SDP cross-attention fusion, FusionVul multimodal vulnerability detection, hierarchical CNN fusion of AST/CFG/DDG, and emerging vision-language model applications. We benchmark performance across datasets, analyze fusion strategy trade-offs, address challenges in modality alignment and missing data, and map the trajectory toward unified multimodal foundation models for software quality assurance.
References
[1] A Cross-Attention Gating Mechanism-Based Multimodal Feature Fusion Method for Software Defect Prediction (GMCA-SDP). (2025). Applied Sciences (MDPI), 15(20), 11259.
[2] FusionVul: A Multimodal Feature Fusion Framework for Source Code Vulnerability Detection. (2025/2026). arXiv:2606.08553.
[3] Defect Detection in Source Code via Multimodal Feature Fusion. (2025). Applied Sciences (MDPI), 15(17), 9358.
[4] Bridging Expert Knowledge with Deep Learning Techniques for Just-In-Time Defect Prediction. (2024). arXiv:2403.11079.
[5] Just-In-Time Software Defect Prediction via Bi-modal Change Representation Learning. (2024). arXiv:2410.12107.
[6] Wang, C., et al. (2022). CODE-MVP: Learning to Represent Source Code from Multiple Views with Contrastive Pre-Training. NAACL 2022 Findings. arXiv:2205.02029.
[7] Guo, D., et al. (2022). UniXcoder: Unified Cross-Modal Pre-training for Code Representation. ACL 2022.
[8] Multi-View Feature Fusion Model for Software Bug Repair Pattern Prediction (PatternNet). (2023). Wuhan University Journal of Natural Sciences, 28(6), 493-508.
[9] Abdu, A., et al. Hierarchical CNN Approach for AST/CFG/DDG Multimodal Software Defect Prediction. Journal of Systems and Software.
[10] Investigation and Research on Several Key Issues of Software Defect Prediction. (2025). IET Software.
[11] Zou, Y., Wang, H., Lv, H., Zhao, S. (2025). Deep Learning-Based Cross-Project Defect Prediction: A Comprehensive Survey. QRS-C 2025.
[12] SWE-Bench Multimodal: Evaluating Language Models on Visual Software Engineering Tasks. (2025). AI Benchmarks Compendium.
[13] ScreenCoder: Advancing Visual-to-Code Generation for Front-End Automation via Modular Multimodal Agents. (2025). arXiv:2507.22827.
[14] Si, C., et al. (2025). Design2Code: Benchmarking Multimodal Code Generation for Automated Front-End Engineering. NAACL 2025.
[15] Giray, G., Bennin, K.E., Köksal, Ö., Babur, Ö., Tekinerdogan, B. (2023). On the Use of Deep Learning in Software Defect Prediction. Journal of Systems and Software, 195, 111537.
[16] Omri, S. & Sinz, C. (2020). Deep Learning for Software Defect Prediction: A Survey. ICSE 2020 Workshops.
How to cite this paper
@article{1719451,
author = {Pooja Ganesh Dhone, Dr. Brijendra Gupta},
title = {Intelligent Software Defect Prediction Using Multimodal Deep Learning Through the Integration of Source Code, Software Metrics and Historical Development Data},
journal = {Iconic Research And Engineering Journals},
year = {2026},
volume = {10},
number = {1},
pages = {226-235},
issn = {2456-8880},
url = {https://www.irejournals.com/formatedpaper/1719451.pdf},
abstract = {Software defect prediction has historically relied on a single view of source code — either handcrafted metrics, token sequences, or structural representations in isolation. Multimodal deep learning addresses the fundamental limitation that no single code view captures the full richness of software artifacts: their syntax, structure, semantics, history, and natural-language context. This report provides a comprehensive survey of multimodal deep learning approaches for software defect prediction, covering the major modalities (lexical/semantic, structural/AST, control-and-data-flow, metric-based, and natural language) and the fusion architectures that combine them — concatenation, attention-gating, cross-attention, and contrastive multi-view learning. We synthesize findings from over 50 recent publications (2020-2026), including GMCA-SDP cross-attention fusion, FusionVul multimodal vulnerability detection, hierarchical CNN fusion of AST/CFG/DDG, and emerging vision-language model applications. We benchmark performance across datasets, analyze fusion strategy trade-offs, address challenges in modality alignment and missing data, and map the trajectory toward unified multimodal foundation models for software quality assurance.},
month = {July},
doi = {https://doi.org/10.64388/IREV10I1-1719451}
}