Banani Roy

Palash R. Roy, Ajmain I. Alam, Farouq Al-omari et al.

Semantic and Cross-language code clone generation may be useful for code reuse, code comprehension, refactoring and benchmarking. OpenAI's GPT model has potential in such clone generation as GPT is used for text generation. When developers copy/paste codes from Stack Overflow (SO) or within a system, there might be inconsistent changes leading to unexpected behaviours. Similarly, if someone possesses a code snippet in a particular programming language but seeks equivalent functionality in a different language, a semantic cross-language code clone generation approach could provide valuable assistance. In this study, using SemanticCloneBench as a vehicle, we evaluated how well the GPT-3 model could help generate semantic and cross-language clone variants for a given fragment.We have comprised a diverse set of code fragments and assessed GPT-3s performance in generating code variants.Through extensive experimentation and analysis, where 9 judges spent 158 hours to validate, we investigate the model's ability to produce accurate and semantically correct variants. Our findings shed light on GPT-3's strengths in code generation, offering insights into the potential applications and challenges of using advanced language models in software development. Our quantitative analysis yields compelling results. In the realm of semantic clones, GPT-3 attains an impressive accuracy of 62.14% and 0.55 BLEU score, achieved through few-shot prompt engineering. Furthermore, the model shines in transcending linguistic confines, boasting an exceptional 91.25% accuracy in generating cross-language clones

Md Nadim, Mohammad Hassan, Ashis Kumar Mandal et al.

Purpose: Quantum computing promises to transform problem-solving across various domains with rapid and practical solutions. Within Software Evolution and Maintenance, Quantum Machine Learning (QML) remains mostly an underexplored domain, particularly in addressing challenges such as detecting buggy software commits from code repositories. Methods: In this study, we investigate the practical application of Quantum Support Vector Classifiers (QSVC) for detecting buggy software commits across 14 open-source software projects with diverse dataset sizes encompassing 30,924 data instances. We compare the QML algorithm PQSVC (Pegasos QSVC) and QSVC against the classical Support Vector Classifier (SVC). Our technique addresses large datasets in QSVC algorithms by dividing them into smaller subsets. We propose and evaluate an aggregation method to combine predictions from these models to detect the entire test dataset. We also introduce an incremental testing methodology to overcome the difficulties of quantum feature mapping during the testing approach. Results: The study shows the effectiveness of QSVC and PQSVC in detecting buggy software commits. The aggregation technique successfully combines predictions from smaller data subsets, enhancing the overall detection accuracy for the entire test dataset. The incremental testing methodology effectively manages the challenges associated with quantum feature mapping during the testing process. Conclusion: We contribute to the advancement of QML algorithms in defect prediction, unveiling the potential for further research in this domain. The specific scenario of the Short-Term Activity Frame (STAF) highlights the early detection of buggy software commits during the initial developmental phases of software systems, particularly when dataset sizes remain insufficient to train machine learning models.

Khairul Alam, Saikat Mondal, Banani Roy

Autonomous coding agents (e.g., OpenAI Codex, Devin, GitHub Copilot) are increasingly used to generate fix-related pull requests (PRs) in real world software repositories. However, their practical effectiveness depends on whether these contributions are accepted and merged by project maintainers. In this paper, we present an empirical study of AI agent involved fix related PRs, examining both their integration outcomes, latency, and the factors that hinder successful merging. We first analyze 8,106 fix related PRs authored by five widely used AI coding agents from the AIDEV POP dataset to quantify the proportions of PRs that are merged, closed without merging, or remain open. We then conduct a manual qualitative analysis of a statistically significant sample of 326 closed but unmerged PRs, spending approximately 100 person hours to construct a structured catalog of 12 failure reasons. Our results indicate that test case failures and prior resolution of the same issues by other PRs are the most common causes of non integration, whereas build or deployment failures are comparatively rare. Overall, our findings expose key limitations of current AI coding agents in real world settings and highlight directions for their further improvement and for more effective human AI collaboration in software maintenance.

Amit Kumar Mondal, Chanchal K. Roy, Kevin A. Schneider et al.

Software architectural changes involve more than one module or component and are complex to analyze compared to local code changes. Development teams aiming to review architectural aspects (design) of a change commit consider many essential scenarios such as access rules and restrictions on usage of program entities across modules. Moreover, design review is essential when proper architectural formulations are paramount for developing and deploying a system. Untangling architectural changes, recovering semantic design, and producing design notes are the crucial tasks of the design review process. To support these tasks, we construct a lightweight tool [4] that can detect and decompose semantic slices of a commit containing architectural instances. A semantic slice consists of a description of relational information of involved modules, their classes, methods and connected modules in a change instance, which is easy to understand to a reviewer. We extract various directory and naming structures (DANS) properties from the source code for developing our tool. Utilizing the DANS properties, our tool first detects architectural change instances based on our defined metric and then decomposes the slices (based on string processing). Our preliminary investigation with ten open-source projects (developed in Java and Kotlin) reveals that the DANS properties produce highly reliable precision and recall (93-100%) for detecting and generating architectural slices. Our proposed tool will serve as the preliminary approach for the semantic design recovery and design summary generation for the project releases.

Avijit Bhattacharjee, Sristy Sumana Nath, Shurui Zhou et al.

The fork-based development mechanism provides the flexibility and the unified processes for software teams to collaborate easily in a distributed setting without too much coordination overhead.Currently, multiple social coding platforms support fork-based development, such as GitHub, GitLab, and Bitbucket. Although these different platforms virtually share the same features, they have different emphasis. As GitHub is the most popular platform and the corresponding data is publicly available, most of the current studies are focusing on GitHub hosted projects. However, we observed anecdote evidences that people are confused about choosing among these platforms, and some projects are migrating from one platform to another, and the reasons behind these activities remain unknown.With the advances of Software Heritage Graph Dataset (SWHGD),we have the opportunity to investigate the forking activities across platforms. In this paper, we conduct an exploratory study on 10popular open-source projects to identify cross-platform forks and investigate the motivation behind. Preliminary result shows that cross-platform forks do exist. For the 10 subject systems in this study, we found 81,357 forks in total among which 179 forks are on GitLab. Based on our qualitative analysis, we found that most of the cross-platform forks that we identified are mirrors of the repositories on another platform, but we still find cases that were created due to preference of using certain functionalities (e.g. Continuous Integration (CI)) supported by different platforms. This study lays the foundation of future research directions, such as understanding the differences between platforms and supporting cross-platform collaboration.

Saumendu Roy, Banani Roy, Chanchal Roy et al.

Machine learning (ML)-based defect prediction models can improve software quality. However, their opaque reasoning creates an HCI challenge because developers struggle to trust models they cannot interpret. Explainable AI (XAI) methods such as LIME, SHAP, and BreakDown aim to provide transparency, but when used together, they often produce conflicting explanations that increase confusion, frustration, and cognitive load. To address this usability challenge, we introduce XMENTOR, a human-centered, rank-aware aggregation method implemented as a VS Code plugin. XMENTOR unifies multiple post-hoc explanations into a single, coherent view by applying adaptive thresholding, rank and sign agreement, and fallback strategies to preserve clarity without overwhelming users. In a user study, nearly 90% of the participants preferred aggregated explanations, citing reduced confusion and stronger support for daily tasks of debugging and review of defects. Our findings show how combining explanations and embedding them into developer workflows can enhance interpretability, usability, and trust.

Ajmain Inqiad Alam, Palash Roy, Chanchal K. Roy et al.

The accelerating adoption of Large Language Models (LLMs) in software engineering (SE) has brought with it a silent crisis: unsustainable computational cost. While these models demonstrate remarkable capabilities in different SE tasks, they are unmanageably large, slow to deploy, memory-intensive, and carbon-heavy. This reality threatens not only the scalability and accessibility of AI-powered SE, but also its long-term environmental sustainability. The research challenge is clear: we must go beyond accuracy and address efficiency and environmental cost as first-class design constraints. To meet this challenge, we introduce Carbon-Taxed Transformers (CTT), a systematic multi-architectural compression principled pipeline ordering inspired by economic carbon taxation principles. Drawing from the economic concept of carbon pricing, CTT operationalizes a computational carbon tax that penalizes architectural inefficiencies and rewards deployment-ready compression. We evaluate CTT across three core SE tasks: code clone detection, code summarization, and code generation, with models spanning encoder-only, encoder-decoder, and decoder-only architecture. Our results show that CTT delivers on inference: (1) up to 49x memory reduction, (2) time reduction up to 8-10x for clone detection, up to 3x for summarization, and 4-7x for generation, (3) up to 81% reduction in CO2 emissions and (4) CTT retains around 98% accuracy on clone detection, around 89% on summarization, and up to 91% (textual metrics) and 68% (pass@1) for generation. Two ablation studies show that pipeline ordering and individual component contributions are both essential, providing empirical justification for CTT's design and effectiveness. This work establishes a viable path toward responsible AI in SE through aggressive yet performance-preserving compression.

Saumendu Roy, Saikat Mondal, Banani Roy et al.

The lack of interpretability is a major barrier that limits the practical usage of AI models. Several eXplainable AI (XAI) techniques (e.g., SHAP, LIME) have been employed to interpret these models' performance. However, users often face challenges when leveraging these techniques in real-world scenarios and thus submit questions in technical Q&A forums like Stack Overflow (SO) to resolve these challenges. We conducted an exploratory study to expose these challenges, their severity, and features that can make XAI techniques more accessible and easier to use. Our contributions to this study are fourfold. First, we manually analyzed 663 SO questions that discussed challenges related to XAI techniques. Our careful investigation produced a catalog of seven challenges (e.g., disagreement issues). We then analyzed their prevalence and found that model integration and disagreement issues emerged as the most prevalent challenges. Second, we attempt to estimate the severity of each XAI challenge by determining the correlation between challenge types and answer metadata (e.g., the presence of accepted answers). Our analysis suggests that model integration issues is the most severe challenge. Third, we attempt to perceive the severity of these challenges based on practitioners' ability to use XAI techniques effectively in their work. Practitioners' responses suggest that disagreement issues most severely affect the use of XAI techniques. Fourth, we seek agreement from practitioners on improvements or features that could make XAI techniques more accessible and user-friendly. The majority of them suggest consistency in explanations and simplified integration. Our study findings might (a) help to enhance the accessibility and usability of XAI and (b) act as the initial benchmark that can inspire future research.

Ashis Kumar Mandal, Md Nadim, Chanchal K. Roy et al.

Predicting software defects early in the development process not only enhances the quality and reliability of the software but also decreases the cost of development. A wide range of machine learning techniques can be employed to create software defect prediction models, but the effectiveness and accuracy of these models are often influenced by the choice of appropriate feature subset. Since finding the optimal feature subset is computationally intensive, heuristic and metaheuristic approaches are commonly employed to identify near-optimal solutions within a reasonable time frame. Recently, the quantum computing paradigm quantum annealing (QA) has been deployed to find solutions to complex optimization problems. This opens up the possibility of addressing the feature subset selection problem with a QA machine. Although several strategies have been proposed for feature subset selection using a QA machine, little exploration has been done regarding the viability of a QA machine for feature subset selection in software defect prediction. This study investigates the potential of D-Wave QA system for this task, where we formulate a mutual information (MI)-based filter approach as an optimization problem and utilize a D-Wave Quantum Processing Unit (QPU) solver as a QA solver for feature subset selection. We evaluate the performance of this approach using multiple software defect datasets from the AEEM, JIRA, and NASA projects. We also utilize a D-Wave classical solver for comparative analysis. Our experimental results demonstrate that QA-based feature subset selection can enhance software defect prediction. Although the D-Wave QPU solver exhibits competitive prediction performance with the classical solver in software defect prediction, it significantly reduces the time required to identify the best feature subset compared to its classical counterpart.

Saikat Mondal, Banani Roy

Software developers often look for solutions to their code-level problems by submitting questions to technical Q&A websites like Stack Overflow (SO). They usually include example code segments with questions to describe the programming issues. SO users prefer to reproduce the reported issues using the given code segments when they attempt to answer the questions. Unfortunately, such code segments could not always reproduce the issues due to several unmet challenges (e.g., external library not found) that might prevent questions from receiving prompt and appropriate solutions. A previous study produced a catalog of potential challenges that hinder the reproducibility of issues reported at SO questions. However, it is unknown how the practitioners (i.e., developers) perceive the challenge catalog. Understanding the developers' perspective is inevitable to introduce interactive tool support that promotes reproducibility. We thus attempt to understand developers' perspectives by surveying 53 users of SO. In particular, we attempt to -- (1) see developers' viewpoints on the agreement to those challenges, (2) find the potential impact of those challenges, (3) see how developers address them, and (4) determine and prioritize tool support needs. Survey results show that about 90% of participants agree to the already exposed challenges. However, they report some additional challenges (e.g., error log missing) that might prevent reproducibility. According to the participants, too short code segment and absence of required Class/Interface/Method from code segments severely prevent reproducibility, followed by missing important part of code. To promote reproducibility, participants strongly recommend introducing tool support that interacts with question submitters with suggestions for improving the code segments if the given code segments fail to reproduce the issues.

Debasish Chakroborti, Banani Roy, Sristy Sumana Nath

To process a large amount of data sequentially and systematically, proper management of workflow components (i.e., modules, data, configurations, associations among ports and links) in a Scientific Workflow Management System (SWfMS) is inevitable. Managing data with provenance in a SWfMS to support reusability of workflows, modules, and data is not a simple task. Handling such components is even more burdensome for frequently assembled and executed complex workflows for investigating large datasets with different technologies (i.e., various learning algorithms or models). However, a great many studies propose various techniques and technologies for managing and recommending services in a SWfMS, but only a very few studies consider the management of data in a SWfMS for efficient storing and facilitating workflow executions. Furthermore, there is no study to inquire about the effectiveness and efficiency of such data management in a SWfMS from a user perspective. In this paper, we present and evaluate a GUI version of such a novel approach of intermediate data management with two use cases (Plant Phenotyping and Bioinformatics). The technique we call GUI-RISPTS (Recommending Intermediate States from Pipelines Considering Tool-States) can facilitate executions of workflows with processed data (i.e., intermediate outcomes of modules in a workflow) and can thus reduce the computational time of some modules in a SWfMS. We integrated GUI-RISPTS with an existing workflow management system called SciWorCS. In SciWorCS, we present an interface that users use for selecting the recommendation of intermediate states (i.e., modules' outcomes). We investigated GUI-RISP's effectiveness from users' perspectives along with measuring its overhead in terms of storage and efficiency in workflow execution.

Golam Mostaeen, Banani Roy, Chanchal Roy et al.

A code clone is a pair of code fragments, within or between software systems that are similar. Since code clones often negatively impact the maintainability of a software system, several code clone detection techniques and tools have been proposed and studied over the last decade. To detect all possible similar source code patterns in general, the clone detection tools work on the syntax level while lacking user-specific preferences. This often means the clones must be manually inspected before analysis in order to remove those false positives from consideration. This manual clone validation effort is very time-consuming and often error-prone, in particular for large-scale clone detection. In this paper, we propose a machine learning approach for automating the validation process. Our machine learning-based approach is used to automatically validate clones without human inspection. Thus the proposed approach can be used to remove the false positive clones from the detection results, automatically evaluate the precision of any clone detectors for any given set of datasets, evaluate existing clone benchmark datasets, or even be used to build new clone benchmarks and datasets with minimum effort. In an experiment with clones detected by several clone detectors in several different software systems, we found our approach has an accuracy of up to 87.4% when compared against the manual validation by multiple expert judges. The proposed method also shows better results in several comparative studies with the existing related approaches for clone classification.

Amit Kumar Mondal, Banani Roy, Chanchal K. Roy et al.

With the rapid advancement of Big Data platforms such as Hadoop, Spark, and Dataflow, many tools are being developed that are intended to provide end users with an interactive environment for large-scale data analysis (e.g., IQmulus). However, there are challenges using these platforms. For example, developers find it difficult to use these platforms when developing interactive and reusable data analytic tools. One approach to better support interactivity and reusability is the use of microlevel modularisation for computation-intensive tasks, which splits data operations into independent, composable modules. However, modularizing data and computation-intensive tasks into independent components differs from traditional programming, e.g., when accessing large scale data, controlling data-flow among components, and structuring computation logic. In this paper, we present a case study on modularizing real world computationintensive tasks that investigates the impact of modularization on processing large scale image data. To that end, we synthesize image data-processing patterns and propose a unified modular model for the effective implementation of computation-intensive tasks on data-parallel frameworks considering reproducibility, reusability, and customization. We present various insights of using the modularity model based on our experimental results from running image processing tasks on Spark and Hadoop clusters.