Call for papers, Special Issue on Intelligent Techniques 2026

Name of the call: Intelligent Techniques for Automated Code Review and Software Quality Evaluation (2026)

Abstract

Automated code review and software quality evaluation are becoming increasingly important in modern software engineering, where codebases evolve rapidly and development cycles continue to shorten. Traditional quality assurance practices often rely on manual inspection, handcrafted rules, and isolated analysis tools, which may struggle to scale across large software ecosystems or adapt to changing coding patterns and development requirements. Recent advances in artificial intelligence have created new opportunities to improve the effectiveness, efficiency, and intelligence of software quality assurance.

A wide range of AI-driven techniques, including Automated Machine Learning (AutoML), large language models (LLMs), program analysis, and multimodal learning, are reshaping how software systems are reviewed, analyzed, and maintained. These techniques support a variety of tasks such as defect prediction, bug localization, vulnerability detection, code smell identification, maintainability assessment, refactoring recommendation, and review assistance. In particular, AutoML enables the automatic construction and optimization of predictive models through neural architecture search, hyperparameter optimization, meta-learning, and automated feature engineering, while LLMs and related foundation models provide strong capabilities in code understanding, reasoning, summarization, and developer interaction.

The combination of these emerging approaches offers new possibilities for intelligent software engineering. By jointly leveraging source code, development history, natural language documentation, execution traces, and developer feedback, AI-driven systems can provide more accurate quality evaluation, more explainable review decisions, and more adaptive support for continuous integration and delivery workflows. At the same time, important challenges remain, including scalability, generalization across projects and languages, interpretability, robustness, trustworthiness, and human-AI collaboration in code review processes.

This thematic collection aims to bring together researchers, practitioners, and industry experts working on intelligent techniques for automated code review and software quality evaluation. It seeks to advance the state of the art in AI-driven software engineering by encouraging original contributions on novel methods, practical systems, empirical studies, and industrial applications.

Topics of interest include, but are not limited to:

  • AutoML-driven models for defect prediction, bug localization, and vulnerability detection
  • Automated neural architecture search for code representation learning
  • Meta-learning and transfer learning for cross-project software quality prediction
  • Large language models for code review, code understanding, and quality assessment
  • AI-driven models for defect prediction, bug localization, and vulnerability detection
  • Surrogate-assisted and performance-prediction methods for accelerating software analytics
  • Integration of LLMs and AutoML for intelligent code understanding
  • Adaptive AutoML frameworks for CI/CD, DevOps, and real-time quality monitoring
  • Intelligent techniques for CI/CD, DevOps, software evolution, technical debt analysis, and refactoring recommendation

Therefore, this thematic collection aligns closely with the scope of Automated Software Engineering in terms of its technical focus, application scenarios, and engineering practices, while also resonating with the ongoing trends toward intelligent, automated, and data-driven software development. The collection aims to provide a high-quality forum for researchers and practitioners to advance the state of the art in AutoML, intelligent code review, and software quality evaluation, thereby fostering the next stage of progress in intelligent software engineering.

This thematic collection welcomes a broad range of submission types, including full research papers, short papers, review articles, vision papers, tool demonstrations, and practice-oriented or industry case studies, to encourage diverse perspectives and contributions from both academia and industry.

Rationale and Timeliness of the Thematic Collection

The diversity and recency of the studies listed above further underscore the timeliness and necessity of this thematic collection. These contributions originate from research groups across the globe and include several works published in 2025, reflecting the rapid growth and strong momentum in AutoML-driven code review and software quality research. The emergence of these cutting-edge studies highlights both the expanding interest in this domain and the current lack of a dedicated venue that brings together advances in AutoML, intelligent code analysis, and automated quality evaluation. By providing a focused platform for these rapidly evolving research directions, this special issue addresses a clear community need and supports the continued development of next-generation intelligent software engineering techniques.

Deadlines

  • CFP Distribution: April, 2026
  • Submission Deadline: October, 2026
  • First Review Decisions: December, 2026
  • Revised Paper Submission: February, 2027
  • Final Acceptance Notification: March, 2027
  • https://2024.ieeewcci.org/special-sessions
  • https://2025.ijcnn.org/authors/special-sessions -https://www.conference2go.com/event/ieee-2025-7th-international-conference-on-data-driven-optimization-of-complex-systems-docs-2025/

Organizers

  • Nan Li is with Institute of Big Data Science and Industry, Shanxi University, and the personal website is https://www.researchgate.net/profile/Nan-Li-234
  • Xinyan Liang is with Institute of Big Data Science and Industry, Shanxi University, and the personal website is https://xinyanliang.github.io/
  • Jun Yu is with Institute of Science and Technology, Niigata University, and the personal website is https://www.eng.niigata-u.ac.jp/~yujun/indexEn.html
  • Wellington P. Santos is with Depto. de Eng. Biomédica, Universidade Federal de Pernambuco, and the personal website is https://www.researchgate.net/profile/Wellington-Dos-Santos https://scholar.google.com/citations?user=7OE2eYEAAAAJ&hl=vi

Contact

  • Nan Li: linan10@sxu.edu.cn
  • Xinyan Liang: liangxinyan@sxu.edu.cn
  • Jun Yu: yujun@ie.niigata-u.ac.jp
  • Wellington P. Santos: wellington.santos@ufpe.br

Organizers Short Biography

  • Nan Li received the B.Sc. degree in software engineering from the North University of China, Taiyuan, China, in 2020. He is currently working toward the Ph.D. degree in software engineering with the College of Software, Northeastern University, Shenyang, China. He interests include computational intelligence and machine learning. He has been selected for the inaugural China Association for Science and Technology (CAST) Young Talent Support Program’s Doctoral Student Special Project. To date, he has published 20 papers in prestigious journals and conferences including ACM CSUR, IJCAI, IEEE TEVC, IEEE TFS, and IEEE TCYB. Among these, 10 papers were authored as first or second author (with the supervisor as first author), including 3 ESI highly cited papers, 1 hot paper, and 3 research frontier papers, garnering over 500 Google citations.

  • Xinyan Liang earned his Ph.D. in Computer Science and Technology from Shanxi University in 2022 and served as a Research Assistant at the University of Hong Kong from 2017 to 2018. Joined the Big Data Science and Industry Research Institute at Shanxi University in August 2022. Current research focuses on multimodal machine learning, signal processing, and their applications in microwave atomic detection. In recent years, he has led five projects: a JKW Key Project, a sub-project/young investigator project under the National Natural Science Foundation of China Major Project, a Shanxi Provincial Major Project, and an interdisciplinary construction project at the university level. He has published over 40 papers in prestigious international journals and conferences, including IEEE TPAMI, IEEE TEVC, IEEE TCSVT, IEEE TFS, ICLR, ICML, AAAI, and IJCAI, with over 20 papers as first/corresponding author.

  • Jun Yu received the bachelor’s degree from Northeastern University, China, in 2014, and the master’s and Ph.D. degrees from Kyushu University, Japan, in 2017 and 2019, respectively. He is currently an Assistant Professor with Niigata University, Japan. His research interests include evolutionary computation, artificial neural networks, and machine learning.

-Wellington P. Santos received the B.Sc. and M.Sc. degrees in Electrical Engineering from the Federal University of Pernambuco, Brazil, in 2001 and 2003, respectively, and the Ph.D. degree in Electrical Engineering from the Federal University of Campina Grande in 2009. He is currently an Associate Professor with tenure at the Department of Biomedical Engineering, Federal University of Pernambuco, where he teaches undergraduate and graduate courses and was one of the founders of the Biomedical Engineering Graduate Program in 2011. He also holds a position in the Graduate Program in Computer Engineering at the Polytechnic School of Pernambuco, University of Pernambuco, since 2009. He leads the Biomedical Computing and Bioengineering axis at the National Institute of Science and Technology for Health Technology (iCEIS). His research lies at the intersection of biomedical engineering and computer science, with focus on digital image processing, pattern recognition, computer vision, evolutionary computation, optimization, computational intelligence, medical imaging, virtual reality, and applications of computing in medicine and biology. He has authored 93 journal papers, 194 conference papers, 28 books, and 74 book chapters, developed 13 open-source software tools, and filed 6 patents. His work has been cited over 4,700 times, and he holds an h-index of 35 (Google Scholar). He is a member of the Brazilian Society of Biomedical Engineering (SBEB), the Brazilian Society of Computational Intelligence (SBIC), the Brazilian Society of Health Informatics (SBIS), and the International Federation for Medical and Biological Engineering (IFMBE).

Potential submissions

  1. Assessing the Use of AutoML for Data-Driven Software Engineering Authors: Fabio Calefato, Luigi Quaranta, Filippo Lanubile, Marcos Kalinowski arXiv: 2307.10774 Summary: This paper presents the first systematic investigation of AutoML tools in software engineering tasks. The authors evaluate 12 AutoML systems across two SE datasets and complement the study with surveys and interviews. Results show that AutoML can match or exceed manually engineered defect prediction pipelines, while still suffering from uneven automation across preprocessing, feature engineering, and pipeline selection. The study highlights both the opportunities and limitations of using AutoML in real SE workflows.

  2. MBL-CPDP: A Multi-objective Bilevel Method for Cross-Project Defect Prediction via AutoML Authors: Jiaxin Chen, Jinliang Ding, Kay Chen Tan, Jiancheng Qian, Ke Li arXiv: 2411.06491 Summary: The authors propose a bilevel multi-objective AutoML framework for cross-project defect prediction. The method jointly optimizes feature selection, transfer adaptation, and classifier choices at the upper level, while fine-tuning hyperparameters at the lower level. Experiments on 20 projects demonstrate significant improvements in robustness and adaptiveness across heterogeneous codebases, showcasing the potential of AutoML in SE predictive analytics.

  3. Deep Learning-Based Code Reviews: A Paradigm Shift or a Double-Edged Sword? Authors: Rosalia Tufano, Alberto Martin-Lopez, Ahmad Tayeb, Ozren Dabić, Sonia Haiduc, Gabriele Bavota arXiv: 2411.11401 Summary: This controlled experiment evaluates whether neural-based automated code review improves developers’ performance. Findings reveal mixed outcomes: reviewers tend to accept most model-suggested issues, but automation does not significantly improve high-severity defect detection. Though helpful in guiding attention, automatic suggestions may also bias reviewers. The work provides nuanced insight into the risks and benefits of ML-driven code review.

  4. Code Review Automation using Retrieval-Augmented Generation Authors: Qianru Meng, Xiao Zhang, Zhaochen Ren, Joost Visser arXiv: 2511.05302 Summary: This paper introduces RARe, a retrieval-augmented pipeline for generating code review comments. The system retrieves relevant historical review snippets and feeds them into an LLM for context-aware generation. Experiments show substantial improvements in BLEU scores and human-judged relevance over both pure retrieval and pure generation baselines. The work demonstrates the power of hybrid RAG methods in automated code review.

  5. Harnessing Large Language Models for Curated Code Reviews Authors: Oussama Ben Sghaier, Martin Weyssow, Houari Sahraoui arXiv: 2502.03425 Summary: The paper focuses on improving the quality of training datasets for code review automation. Using LLMs, the authors curate and refine existing large-scale code review datasets by filtering noisy comments and rewriting weak samples. The resulting curated dataset improves downstream comment-generation model performance, highlighting the importance of data quality for ML-based code review research.

  6. Too Noisy to Learn: Enhancing Data Quality for Code Review Comment Generation Authors: Chunhua Liu, Hong Yi Lin, Patanamon Thongtanunam arXiv: 2502.02757 Summary: This study examines the prevalence and negative impact of noisy review comments in open-source datasets. It employs LLM-based semantic filtering to isolate actionable and high-quality comments. Retraining models on the cleaned dataset significantly improves comment informativeness and coherence, underscoring that dataset noise remains a core obstacle to automated code review.

  7. Towards Practical Defect-Focused Automated Code Review Authors: Junyi Lu, Lili Jiang, Xiaojia Li, Jianbing Fang, Fengjun Zhang, Li Yang, Chun Zuo arXiv: 2505.17928 Summary: This industrial case study redesigns automated code review around real defect discovery rather than generic comment generation. The authors propose a multi-role LLM workflow, code slicing techniques, and filtering modules tailored for a massive C++ codebase in a commercial environment. The approach significantly improves defect coverage and relevance, illustrating how practical constraints reshape ACR system design.

  8. Automated Code Review Using Large Language Models with Symbolic Reasoning Authors: Busra Icoz, Goksel Biricik arXiv: 2507.18476 Summary: The paper proposes a hybrid LLM–symbolic reasoning framework for automated code review. By integrating symbolic analysis with neural code models such as CodeT5 and CodeBERT, the system improves error detection accuracy and reduces false positives. Results highlight the benefits of combining neural and symbolic signals for safety-critical code analysis.

  9. Evaluating Large Language Models for Code Review Authors: Umut Cihan, Arda İçöz, Vahid Haratian, Eray Tüzün arXiv: 2505.20206 Summary: This evaluation study tests GPT-4o and Gemini 2.0 Flash on correctness checking and suggestion generation for code review tasks. LLMs show strong performance when problem descriptions are provided but inconsistent behavior across different code types. The authors advocate for hybrid workflows where LLMs support rather than replace human reviewers.

  10. Automated Code Review in Practice Authors: Umut Cihan, Vahid Haratian, Arda İçöz, Mert Kaan Gül, Ömercan Devran, Emircan Furkan Bayendur, Baykal Mehmet Uçar, Eray Tüzün arXiv: 2412.18531 Summary: This real-world deployment study evaluates an LLM-powered automated code review tool (based on Qodo PR Agent) across ten production projects. Results show that more than two-thirds of automatically generated comments are eventually addressed, demonstrating practical value. However, the study also identifies cost, delay, and noise issues, offering insights into real-world challenges of integrating ACR tools into development pipelines.