Science AI Journal
Computer Science · Research topic

Open research questions in Software Engineering Research

371 gap statements mined from Software Engineering Research papers in our 7.9M-paper local library, which holds 2,133 papers on the topic — drawn mostly from each paper's own stated research gap, future-work, challenge and limitation notes, and its abstract. The ones listed below are a selection still marked open; each names the study that raised it, with a DOI link where the paper has one.

Choosing where to publish on Software Engineering Research? See the ranked Computer Science journals.

What the literature leaves open

  • The lack of a comprehensive evaluation framework to assess the quality of HTML code generated by Generative AI models. The need to identify the limitations and strengths of current AI tools in HTML development. The gap in understanding the capabilities and limitations of Generative AI models in HTML generation.

    Evaluating Generative AI for HTML Development · 2025 · DOI

  • Although recent LLM-based FL methods have improved localization accuracy, their impact on agentic repair remains unclear.

    From Finding to Fixing: How Useful is Fault Localization for Agentic Program Repair? · 2026

  • Entity--Boundary--Control (EBC) separates domain information, external interaction, and workflow coordination, but its proportions across diverse software remain unclear.

    The EBC Proportionality Law: A Recurring Architectural-Role Pattern Across Java Software · 2026

  • Large language models can detect vulnerable code, yet the form of vulnerability encoding in their representations remains unclear.

    When Models Know but Can't Linearly Say: Relational Encoding Limits Linear Vulnerability Classification in Code LLMs · 2026

  • However, such experience is diverse and evolving, so it remains unclear how memory systems organize it and thereby affect code agents.

    MemDev-Bench: Does Memory Help Code Agents?Benchmarking Memory Systems over Real Repository Histories · 2026

  • These observations warrant further evaluation of transparent invocation control, with conclusions limited to the tested tasks, models, and sampling conditions.

    Cost-Aware Python Code Repair via Budget-Threshold LLM Optimization · 2026 · DOI

  • Although static program analysis methods are frequently employed to enhance software quality, their efficiency in commercial settings is limited by their high false positive rate.

    An empirical application of user-guided program analysis · 2024 · DOI

  • Specifically, future work should focus on developing verification methods independent of complete test suites, constructing architecture-aware models that can understand project-level dependencies, and incorporating non-functional goals such as efficiency and security into migration constraints.

    A Multi-Perspective Investigation into Code Migration for Large Language Models · 2026 · DOI

  • First, in view of the over-reliance of code migration verification on existing test cases, future research should explore verification methods in the absence of testing environments.

    A Multi-Perspective Investigation into Code Migration for Large Language Models · 2026 · DOI

  • The lack of systematic surveys on LLM-based clone detection. The limited generalization ability of traditional fine-tuned models.

    A Comprehensive Study of LLM-Based Code Clone Detection · 2026 · DOI

  • AI systems remain limited in their ability to support complex architectural reasoning. AI-generated code may not reflect unique coding conventions and architectural constraints used within organizations. Evaluating AI-generated code requires assessing maintainability, security practices, and architectural consistency.

    The Metamorphosis of Code: How Artificial Intelligence is Reshaping Software Architecture · 2026 · DOI

  • Future research should focus on improving the ability of AI systems to support complex architectural reasoning. Developing more robust testing methodologies is necessary to address the fragility of AI-generated test code.

    The Metamorphosis of Code: How Artificial Intelligence is Reshaping Software Architecture · 2026 · DOI

  • This study addresses a significant gap in the literature by proposing a proactive approach for software module dependency prediction.

    From Reactive to Proactive: Graph Convolutional Networks for Future Software Module Coupling Prediction · 2026 · DOI

  • The challenge of detecting multiple code smells that may affect the same software component. The difficulty of capturing the structural characteristics of source code required for code smell classification. The need to prevent information leakage between the training and test sets.

    Evaluating Embedding Representations for Multiclass Code Smell Detection: A Comparative Study of CodeBERT and General-Purpose Embeddings · 2026 · DOI

  • Lack of adaptability and extensibility in traditional SDK generation tools. High cognitive load in monolithic LLM approaches.

    Reducing Cognitive Load in Multi-Language SDK Generation Using Hybrid AI-Orchestrated Architectures · 2026 · DOI

  • The lack of domain-specific knowledge can limit the effectiveness of LLMs in specific scenarios. The high computational resource requirements associated with fine-tuning LLMs can limit their use.

    Software defect detection using large language models: a literature review · 2026 · DOI

  • Future research can focus on exploring other feature selection techniques and data balancing methods. The study suggests that further evaluation of the proposed methods on larger datasets is necessary.

    Evaluating the indispensable aspect of data ımbalance in software fault prediction with feature selection and variants of machine learning and ensemble techniques · 2026 · DOI

  • Cloud-based deployment is mentioned as a scalability enhancement, but specific architectural decisions (containerization technology, distributed execution strategy, resource allocation policies) and performance comparisons between local sandbox and cloud-based execution are not detailed.

    AI-Powered Code Helper for Intelligent Code Analysis, Debugging, and Multi-Language Execution · 2026 · DOI

  • The results show significant performance variation across different CodeLMs, with small CodeLMs (the parameter size is 220M or less) tending to achieve better performance than large CodeLMs within our experimental setting.

    Fine-Tuning Code Language Models to Detect Cross-Language Bugs · 2026 · DOI

  • Implementation errors in the active learning algorithms and acquisition functions, - Selection of hyperparameters for Gaussian Process Models and Tree of Parzen Estimators may introduce subtle biases, - Findings may not generalize to tasks involving vastly different domains, - Limited diversity of real-world SE problems in the datasets, - Limited to 49 SE…

    Can Large Language Models Improve SE Active Learning via Warm Starts? · 2026 · DOI

  • Expand the dataset to encompass a broader variety of SE optimization problems, - Explore methods that only need a few dozen examples, - Investigate the use of LLMs in high-dimensional problems, - Replication by external researchers…

    Can Large Language Models Improve SE Active Learning via Warm Starts? · 2026 · DOI

  • Extend the framework with computational efficiency measures, - Compare AI-generated HTML to human-coded baselines, - Assess real-world viability while maintaining quality…

    Evaluating Generative AI for HTML Development · 2025 · DOI

  • Developing more robust and consistent LLMs. Improving the recall rate for Type-4 clone detection. Exploring new technical ways for code clone detection.

    A Comprehensive Study of LLM-Based Code Clone Detection · 2026 · DOI

  • Exploring the application of CodeBERT to other code analysis tasks, - Investigating the use of other embedding models for code smell detection, - Analyzing the performance of CodeBERT on larger and more diverse…

    Evaluating Embedding Representations for Multiclass Code Smell Detection: A Comparative Study of CodeBERT and General-Purpose Embeddings · 2026 · DOI

  • Investigating the application of the proposed hybrid architecture to other code generation tasks. Exploring the use of different LLM configurations and system complexities to further improve the effectiveness of the proposed architecture.

    Reducing Cognitive Load in Multi-Language SDK Generation Using Hybrid AI-Orchestrated Architectures · 2026 · DOI

Most-cited papers in Software Engineering Research

Most recent work

Find a gap in your own Software Engineering Research sub-topic

This page shows what the Software Engineering Research literature already flags as unresolved. To narrow it to your specific question, search the Research Gap Finder: the search is free with a free account and lists the papers closest to your topic first. Unlocking that topic (50 credits, charged once) fills the comparison table from our 7.9M-paper local library and writes the gaps from its rows.

Open the Research Gap Finder →

Related topics in Computer Science

Online Learning and Analytics
1,253 open questions
Advanced Neural Network Applications
1,236 open questions
AI in Service Interactions
1,161 open questions
Virtual Reality Applications and Impacts
1,060 open questions
Computational Drug Discovery Methods
1,019 open questions
Quantum Computing Algorithms and Architecture
877 open questions
AI in cancer detection
835 open questions
Topic Modeling
780 open questions
Open questions raised by papers listed on this page include: (1) In view of the over-reliance of code migration verification on existing test cases, future research should explore verification methods in the absence of testing environments (“A Multi-Perspective Investigation into Code Migration for Large Language Models”, 2026); (2) The lack of systematic surveys on LLM-based clone detection (“A Comprehensive Study of LLM-Based Code Clone Detection”, 2026); (3) AI systems remain limited in their ability to support complex architectural reasoning (“The Metamorphosis of Code: How Artificial Intelligence is Reshaping Software…”, 2026); (4) Lack of adaptability and extensibility in traditional SDK generation tools (“Reducing Cognitive Load in Multi-Language SDK Generation Using Hybrid…”, 2026); (5) The lack of domain-specific knowledge can limit the effectiveness of LLMs in specific scenarios (“Software defect detection using large language models: a literature review”, 2026). 371 gap statements have been mined from Software Engineering Research papers in our 7.9M-paper local library, which holds 2,133 papers on the topic; the gaps come from whichever of those papers state one. They are mostly the research gaps the authors state and the papers' abstracts, plus future-work, limitations and challenges passages. The ones listed below are a selection still marked open; each names the study that raised it, with a DOI link where the paper has one, so you can read the original claim in context.

Tools for your next paper

Gaps & Literature SearchFind open questions in any field — 1,100,000+ open gaps mined from our own library.Pre-CheckIs your paper ready? Acceptance odds for top, mid and open/emerging journals, in seconds.Journal FinderA ranked shortlist of target journals from a 17,500-venue index.Duplicate Publication CheckerCheck for prior publication across 7 sources and our local library.Citation GeneratorFreeA DOI or title → APA, MLA, Vancouver, Chicago, BibTeX, RIS. No signup.Graphical Abstract MakerFreeTurn your findings into a colorblind-safe graphical abstract.

How Pre-Check worksHow AI Review works

Compare the category — Honest roundups of the AI research tools, ours listed alongside the alternatives.

Command palette

Jump anywhere, run any action.