A Systematic Literature Review on the Use of Deep Learning in Software Engineering Research
The DOI matches its Crossref record; 207 of the 207 references it cites were found in a registry or an index.
Abstract
An increasingly popular set of techniques adopted by software engineering (SE) researchers to automate development tasks are those rooted in the concept of Deep Learning (DL). The popularity of such techniques largely stems from their automated feature engineering capabilities, which aid in modeling software artifacts. However, due to the rapid pace at which DL techniques have been adopted, it is difficult to distill the current successes, failures, and opportunities of the current research landscape. In an effort to bring clarity to this cross-cutting area of work, from its modern inception to the present, this article presents a systematic literature review of research at the intersection of SE & DL. The review canvasses work appearing in the most prominent SE and DL conferences and journals and spans 128 papers across 23 unique SE tasks. We center our analysis around the components of learning , a set of principles that governs the application of machine learning techniques (ML) to a given problem domain, discussing several aspects of the surveyed work at a granular level. The end result of our analysis is a research roadmap that both delineates the foundations of DL techniques applied to SE research and highlights likely areas of fertile exploration for the future.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Deep learning models accelerate and enhance the accuracy of complex coding tasks [122, 214, 234].”
From A 2030 Roadmap for Software Engineering (2025) · cites this paper as [214]
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Artificial intelligence can improve efficiency and reduce frustration by fixing errors, recommending best practices, and generating documentation [214].”
From A 2030 Roadmap for Software Engineering (2025) · cites this paper as [214]
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
ImageNet classification with deep convolutional neural networks
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017) · Communications of the ACM · doi:10.1145/3065386 · Cited by 112,155 · cited 1×
“rprisingly, there was very little work done between the years of 2009 and 2014. However, even after the popularization of DL techniques brought about by results achieved by approaches such as AlexNet [73], it took the SE community 3 years to begin exploring the use of DL techniques for SE tasks. This also coincides wi…”
MethodAdam: A Method for Stochastic Optimization
Kingma, D. P., & Ba, J. (2014) · arXiv · doi:10.48550/arxiv.1412.6980 · Cited by 98,946 · cited 1×
“It can increase the rate at which the learning rate is optimized [88].”
BackgroundBERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2018) · arXiv · doi:10.48550/arxiv.1810.04805 · Cited by 73,173
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2019) · Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers) · doi:10.18653/v1/n19-1423 · Cited by 54,325
(2014) · Cited by 33,419
Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
Ioffe, S., & Szegedy, C. (2015) · arXiv · doi:10.48550/arxiv.1502.03167 · Cited by 27,818
(2014) · Cited by 15,747
Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift
Ioffe, S. (2024) · arXiv (Cornell University) · doi:10.57702/o9raffed · Cited by 14,214
(2016) · Cited by 11,679
(2010) · Cited by 8,261
Research works printed but not linked: 7
- [2] [n.d.]. Literature Reviews on Applying Artificial Intelligence/Machine Learning to Software Engineering Research Problems: Preliminary. http://ceur-ws.org/Vol-2506/Paper5-seed2019.pdf -
- [3] [2019. ACM Artifact Review Policies https://www.acm.org/publications/policies/artifact-review-badging.](https://www.acm.org/publications/policies/artifact-review-badging) -
- [16] Tal Ben-Nun, Alice Shoshana Jakobovits, and Torsten Hoefler. 2018. Neural Code Comprehension: A Learnable Representation of Code Semantics. In Advances in Neural Information Processing Systems 31: Annual Conference on Neural Information Processing Systems 2018, NeurIPS 2018, December 3-8, 2018, Montréal, Canada , Samy Bengio, Hanna M. Wallach, Hugo Larochelle, Kristen Grauman, Nicolò Cesa-Bianchi, and Roman Garnett (Eds.). 3589-3601. https://proceedings.neurips.cc/paper/2018/hash/17c3433fecc21b57000debdf7ad5c930-Abstract.html -
- [42] Jacob Devlin, Rudy Bunel, Rishabh Singh, Matthew J. Hausknecht, and Pushmeet Kohli. 2017. Neural Program Meta-Induction. In Advances in Neural Information Processing Systems 30: Annual Conference on Neural Information Processing Systems 2017, December 4-9, 2017, Long Beach, CA, USA , Isabelle Guyon, Ulrike von Luxburg, Samy Bengio, Hanna M. Wallach, Rob Fergus, S. V. N. Vishwanathan, and Roman Garnett (Eds.). 2080-2088. https: //proceedings.neurips.cc/paper/2017/hash/3bf55bbad370a8fcad1d09b005e278c2-Abstract.html -
- [45] Pedro Domingos. 1998. Occam's Two Razors: The Sharp and the Blunt. In Proceedings of the Fourth International Conference on Knowledge Discovery and Data Mining (New York, NY) (KDD'98) . AAAI Press, 37-43. -
- [51] F. Falcini, G. Lami, and A. Mitidieri. 2017. Yet Another Challenge for the Automotive Software: Deep Learning. IEEE Software (2017), 1-1. https://doi.org/10.1109/MS.2017.265101102 - · doi:10.1109/ms.2017.265101102
- [142] J. Sayyad Shirabad and T.J. Menzies. 2005. The PROMISE Repository of Software Engineering Databases. School of Information Technology and Engineering, University of Ottawa, Canada. http://promise.site.uottawa.ca/SERepository -
Cited by
Items 101–139 of 139
Sonawane, S., & Puthran, S. (2024) · Innovations in Systems and Software Engineering · doi:10.1007/s11334-024-00592-z · Cited by 4
RobFace: A Test Suite for Efficient Robustness Evaluation of Face Recognition Systems
Zhang, R., & Sun, J. (2025) · IEEE Transactions on Reliability · doi:10.1109/tr.2025.3554575 · Cited by 2
“This work is also related to works on analyzing neural networks in general [87], [88], [89], [90], [91], [92].”
Weyssow, M., Zhou, X., Kim, K., Lo, D., & Sahraoui, H. A. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2305.04106 · Cited by 1
“In this perspective, previous works [11, 59] have primarily focused on stationary settings, neglecting the practical need for models to adapt to changing environments and data over time.”
BackgroundGu, J., & Gall, H. C. (2023) · ACM Joint European Software Engineering Conference and Symposium on the Foundations of Software Engineering (ESEC/FSE) · doi:10.1145/3611643.3613076 · Cited by 2
“To promote the productivity of software development, using deep learning to assist related activities is a lasting research trend [16, 36].”
BackgroundUnderstanding, Analysis, and Handling of Software Architecture Erosion
Li, R. (2023) · University of Groningen research database (University of Groningen / Centre for Information Technology) · doi:10.33612/diss.833424077 · Cited by 2
(Deep) Learning of Android Access Control Recommendation from Static Execution Paths
Vagavolu, D., Aafer, Y., & Nagappan, M. (2024) · IEEE European Symposium on Security and Privacy · doi:10.1109/eurosp60621.2024.00047
“Deep Learning for Software Engineering In literature, several software engineering tasks have adapted machine learning and deep learning approaches to achieve state-of-the-art performance for their respective tasks [14], [33].”
BackgroundOn the Generalizability of Transformer Models to Code Completions of Different Lengths
Cooper, N., Tufano, R., Bavota, G., & Poshyvanyk, D. (2024) · IEEE International Conference on Software Maintenance (ICSM) · doi:10.1109/icsme58944.2024.00042 · Cited by 1
“Large Language Models (LLMs) for code have achieved state-of-the-art results across a variety of software engineering (SE) tasks such as code completion [1], [2], code reviews [3], code summarization [4], program repair [5], [6], testing [7], [8], and others [9], [10], [11].”
Self-citationBludo, G. Y., & Buabeng-Andoh, F. (2025) · Inproforum ... · doi:10.32725/978-80-7694-102-1.09
Cui, X., & Wang, J. (2024) · International Conference on Educational Technology Management (ICETM) · doi:10.1145/3711403.3711469
“The integration process of scattered knowledge points in software engineering teaching involves label segmentation to identify key feature K , using deep learning to generate text vectors, and then training a single category SVDD model to optimize the hypersphere [6].”
Enhancing Human Diseases Prediction with Nail Images Using VGG16
Marulkar, S., Narain, B., & Mente, R. (2025) · Springer eBooks · doi:10.1007/978-981-97-9006-7_6
Security Mechanism Based Automotive Software Update Management Strategy
Li, M., Ye, Y., Jiang, L., Su, Y., & Li, H. (2024) · International Conference on Big Data, Artificial Intelligence and Risk Management (ICBAR) · doi:10.1145/3718751.3718844
“Watson C analyzed the components of learning and discussed several aspects of survey work at the granularity level [13].”
Das, S., Verma, R., Sarkar, A., Gupta, S., Thakur, S., Padnekar, M., Jain, S., Keshari, A., Kumari, P., & Banerjee, S. (2025) · SSRN Electronic Journal · doi:10.2139/ssrn.5422006
A Survey of Learning-based Method Name Prediction
Qian, H., Liu, W., Xu, T., Yin, J., Feng, X., Sun, W., Ding, Z., Chen, Y., Miao, Y., Li, J., et al. (2025) · ACM Transactions on Software Engineering and Methodology · doi:10.1145/3771999
Breaking Bottlenecks in Large Language Model Inference With Adaptive Cache Management
Mastropaolo, A. (2025) · Computer · doi:10.1109/mc.2025.3604189
Zhu, Z., Peng, L., Wang, Y., Li, Y., & Long, X. (2025) · ACM International Conference on Information and Knowledge Management (CIKM) · doi:10.1145/3746252.3761420
SRVul: A High-Quality Self-Restrained Vulnerable Code Dataset for Vulnerability Detection
Huang, H., Zou, F., Gui, J., Zheng, T., & Wu, Y. (2025) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn64981.2025.11229257
“Constructing such a dataset is challenging [10].”
Dimitriu, C., Carcea, D., Carcea, R., Iaţu, A., & Iftene, A. (2025) · International Conference on Innovations in Intelligent Systems and Applications (INISTA) · doi:10.1109/inista68122.2025.11249671
“From elaborate prototyping to effective coding, many current products claim to boost productivity and brainstorming [3], [10]–[12].”
Abbasi, M. S., Munir, B., Jayaram, V., & Rivas, P. (2025) · IEEE Access · doi:10.1109/access.2025.3609282
“One of the significant challenges in deep learning is overfitting [62], [63], which occurs when a model learns the training data magnificently, including noise and irrelevant details, instead of concentrating on the target premise [63].”
Lindriany, J., Suyatno, Hartanto, D., Putri, M. D. F. A., & Putri, P. A. A. (2025) · Journal of Pedagogy and Education Science · doi:10.56741/iistr.jpes.001137
“…from existing patterns, has the potential to be developed as a supporting instrument for religious moderation through various applications such as sentiment analysis of religious content, detection of religion-based hatred, and personalization of moderate religious education content [26], [27].”
Mapping the landscape of deep learning in software testing: a bibliometric analysis
Raharjana, I. K., Husna, O. I., Hariyanti, E., & Fauzi, S. S. M. (2025) · Discover Artificial Intelligence · doi:10.1007/s44163-025-00596-z · Cited by 1
“DL is used in software engineering to automate feature extraction, predict defects, generate code recommendations, and support testing tasks [7].”
Jian, C., Gao, X., & Ren, J. (2025) · IEEE Advanced Information Management, Communicates, Electronic and Automation Control Conference (IMCEC) · doi:10.1109/imcec66174.2025.11332063
“…and Watson et al. specifically investigated the application of deep learning in software analyzing the practical implementation of deep learning models (such as recurrent neural networks and convolutional neural networks) in tasks like code representation learning and test case generation[8, 9].”
Can Large Language Models Improve SE Active Learning via Warm Starts?
Senthilkumar, L., & Menzies, T. (2026) · ACM Transactions on Software Engineering and Methodology · doi:10.1145/3796511 · Cited by 12
Abboush, M., Ouarrad, H., & Rausch, A. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6211206
Abboush, M., & Rausch, A. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6226241
Zhao, Y., Zeng, Z., Huang, Z., & Gong, L. (2025) · Asia-Pacific Software Engineering Conference (APSEC) · doi:10.1109/apsec66846.2025.00053
Abboush, M., Ouarrad, H., & Rausch, A. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6300884
The Digital Scholar: Artificial Intelligence in Social Work Research
Rai, A., Ravi, K. E., Singer, J. B., Sanchez-Gomez, A., Thiruvathukal, G. K., & Pautsch, E. (2026) · Families in Society The Journal of Contemporary Social Services · doi:10.1177/10443894251407365
Durmaz, A., & Dalveren, G. G. M. (2026) · Turkish Journal of Science and Technology · doi:10.55525/tjst.1759498
Bao, Q. (2026) · Lecture notes in networks and systems · doi:10.1007/978-3-032-14186-6_37
Dhawan, R., & Dhawan, M. (2026) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2026.1776546 · Cited by 4
“Recent surveys document applications in defect prediction, code generation, program repair, test generation, code review automation, effort estimation, and software analytics (Zhang et al., 2013; Watson et al., 2022).”
Bai, Y., Mu, W., & Song, C. (2026) · Computers Environment and Urban Systems · doi:10.1016/j.compenvurbsys.2026.102438
Dai, P., Ma, X., Huang, X., Wang, T., Zhao, Y., & Zhao, C. (2026) · Software Quality Journal · doi:10.1007/s11219-026-09755-6
Combining Textual Embeddings and Graph Neural Networks for Agile Software Effort Estimation
Bitencourt, C., Turine, M. A. S., Borges, V. A., & Nogueira, B. M. (2026) · ACM Symposium on Applied Computing · doi:10.1145/3748522.3780007
Gao, S., Chen, Z., Xie, X., Huang, W. W., CHEN, W., Chen, W., & Huang, Y. (2026) · Environmental Modelling & Software · doi:10.1016/j.envsoft.2026.107101
Fusing semantics and graph neural networks: A multi-dimensional survey and technological evolution
Zhang, Y. L., & Liu, N. (2026) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2026.115584
EchoGuard: ML based Early Warning System for Silent Software Project Failures
Deekshitha, S., Kannan, A., Daisy, G., & Hemalatha, S. (2026) · International Conference on Computer Networks, Big Data and IoT (ICCBI) · doi:10.1109/iccbi68589.2026.11619861
“In software, researchers have tried to apply similar logic using deep learning models like LSTMs to look at code commits [4].”
Closing the Loop: Program-Interoperable Large Language Models Fuzzing for JavaScript Engines
Mastropaolo, A., & Cheng, Z. (2026) · Computer · doi:10.1109/mc.2026.3701431
Hasan, I. H., & Potrus, M. Y. (2026) · Automated Software Engineering · doi:10.1007/s10515-026-00658-1
“GANs have been applied to synthetic test-data generation for privacy-sensitive domains and numerical inputs (Guo et al. 2022; Watson et al. 2022).”
Tuama, A. Y., Hasan, T. M., & Shakir, A. C. (2026) · Journal of Information Security and Applications · doi:10.1016/j.jisa.2026.104608
Metadata
- Authors
- Watson, C., Cooper, N., Palacio, D. N., Moran, K., & Poshyvanyk, D.
- Affiliations
- Washington and Lee University (US)
- William & Mary (US)
- George Mason University (US)
- Venue
- ACM Transactions on Software Engineering and Methodology
- Published
- 2022-03-04
- Volume
- 31
- Issue
- 2
- Pages
- 1-58
- ISSN
- 1049-331X (ISSN-L)
- 1557-7392
- OpenAlex
- W3084812981
- Cited by
- 137
- References
- 204
- Citations by year
- 2026: 21 · 2025: 39 · 2024: 37 · 2023: 30 · 2022: 8
- FWCI
- 3.60
- Type
- journal article
- Language
- en
- Open access
- closed
- Locations
- ACM Transactions on Software Engineering and Methodology (publishedVersion)
- Funders
- NSF (CCF-1927679, CCF-1955853, CCF-2007246, and CCF-1815186)
- Topics
- Physical Sciences › Computer Science › Information Systems › Software Engineering Research
- Physical Sciences › Computer Science › Information Systems › Software Engineering Techniques and Practices
- Physical Sciences › Computer Science › Software › Advanced Software Engineering Methodologies
- Keywords
- deep learning, software engineering research, software development automation, automated feature engineering, machine learning in software engineering, research roadmap, software engineering tasks
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.