Reproduce “Deep Learning-based Software Engineering: Progress, Challenges, and Opportunities” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Xiangping Chen, Xing Hu, Yuan Huang, He Jiang, Weixing Ji, Yanjie Jiang, Yanyan Jiang, Bo Liu and 18 more
A task-oriented survey analyzes the progress, challenges, and opportunities of applying deep learning techniques across twelve major software engineering subareas, identifying widespread adoption and performance improvements while highlighting persistent issues in data quality and model interpretability.
Key insights
- Deep learning has achieved widespread adoption and demonstrated significant performance improvements over traditional methods across numerous software engineering tasks.
- The acquisition of high-quality, large-scale, and representative training data remains a pervasive and critical challenge across nearly all DL4SE applications.
- Interpretability (the "black-box" problem) and effective generalization across different programming languages are fundamental limitations hindering trust and broader adoption of DL models in SE.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.