Science of Computer Programming最新文献

{"title":"What does AI need to know to drive: Testing relevance of knowledge","authors":"Dominik Grundt,&nbsp;Astrid Rakow,&nbsp;Philipp Borchers,&nbsp;Eike Möhlmann","doi":"10.1016/j.scico.2025.103297","DOIUrl":"10.1016/j.scico.2025.103297","url":null,"abstract":"<div><div>Artificial Intelligence (AI) plays an important role in managing the complexity of automated driving. Nonetheless, training and ensuring the safety of AI is challenging. The safe generalization from a known to an unknown situation remains an unsolved problem. Infusing knowledge into AI driving functions seems a promising approach to address generalization, development costs, and training efficiency. We reason that ascertaining the relevance of infused knowledge provides a strong indication of the correct execution of previous development phases of knowledge infusion. As a causal reason for AI performance, relevant knowledge is important for explaining AI behavior. This paper defines a novel notion of <em>relevant knowledge</em> in knowledge-infused AI and for requirements satisfaction in traffic scenarios. We present a scenario-based testing procedure that not only checks whether a knowledge-infused AI model satisfies a given requirement <span>R</span> but also provides statements on the relevance of infused knowledge. Finally, we describe a systematic method for generating abstract knowledge scenarios to enable an efficient application of our relevance testing procedure.</div></div>","PeriodicalId":49561,"journal":{"name":"Science of Computer Programming","volume":"244 ","pages":"Article 103297"},"PeriodicalIF":1.5,"publicationDate":"2025-03-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143551680","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Human-in-the-loop online just-in-time software defect prediction: What have we achieved and what do we still miss?","authors":"Xutong Liu ,&nbsp;Yufei Zhou ,&nbsp;Yutian Tang ,&nbsp;Junyan Qian ,&nbsp;Yuming Zhou","doi":"10.1016/j.scico.2025.103296","DOIUrl":"10.1016/j.scico.2025.103296","url":null,"abstract":"<div><div><strong>Background.</strong> The Online Just-In-Time Software Defect Prediction (O-JIT-SDP) employs an online model to predict whether a new software change will introduce a bug. Previous studies have neglected to consider the interaction between Software Quality Assurance (SQA) personnel and the model, potentially missing opportunities to refine prediction accuracy through human feedback. <strong>Problem.</strong> A recent study introduced the first Human-In-The-Loop (HITL) O-JIT-SDP framework called HumLa, integrating SQA staff feedback without accounting for inspection time to boost the prediction performance of O-JIT-SDP. However, upon a thorough revisit of HumLa, we find that while certain aspects of the HITL O-JIT-SDP system appear feasible in ideal conditions, they prove impractical in real-world context. <strong>Objective.</strong> We aim to reformulate HITL O-JIT-SDP, which are crucial yet absent for practical application. <strong>Method.</strong> We propose four crucial enhancements to facilitate practical application of HITL O-JIT-SDP. First, we advocate for the use of observed labels rather than ground-truth labels to evaluate online classifiers in real-world settings. Second, we suggest refraining from utilizing the entire data stream for normalizing features of each new instance, as was done in HumLa. Third, we propose incorporating non-zero SQA inspection time into the formulation of HITL O-JIT-SDP. Fourth, we introduce real-time statistical classifier comparison into the HITL system. <strong>Result.</strong> Our replication uncovers that the performance evaluation of HumLa under a practical scenario significantly deviate from the originally reported performance under an ideal experimental scenario, potentially diminishing the promise of HITL O-JIT-SDP. Furthermore, with our enhanced HITL O-JIT-SDP framework, we revisit a fundamental question in O-JIT-SDP: the benefits of HITL integration. Our experimental findings demonstrate that HITL not only enhances O-JIT-SDP when SQA feedback surpasses Bug-Fixing Commit (BFC) feedback (by providing training commits with superior label quality in less time) but also improves O-JIT-SDP even when SQA feedback delay equals that of BFC feedback (by consistently delivering training commits with improved label quality). The real-time statistical analysis reveals that HITL approaches generally outperform non-HITL O-JIT-SDP approaches with a statistically significant margin. <strong>Conclusion.</strong> Our work bolsters model evaluation credibility and holds the potential to substantially enhance the value of HITL O-JIT-SDP for industrial applications.</div></div>","PeriodicalId":49561,"journal":{"name":"Science of Computer Programming","volume":"244 ","pages":"Article 103296"},"PeriodicalIF":1.5,"publicationDate":"2025-03-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143551678","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Model checking and verification of a rail-side protection system","authors":"Fatih Sarikoc","doi":"10.1016/j.scico.2025.103286","DOIUrl":"10.1016/j.scico.2025.103286","url":null,"abstract":"<div><div>Ensuring safe operation of level crossings is crucial in preventing fatalities and injuries at rail-road intersections. This study presents a Petri Net model of a level-crossing system that includes an obstacle detection system and dedicated protection signal to enhance safety. The model was analyzed using linear temporal logic and computation tree logic, with formal verification performed via TINA and TAPAAL tools. As a result of the model-checking experiments, the proposed Petri Net model successfully complies with the given safety criteria. Additionally, Fault Tree Analysis (FTA) was employed to systematically assess system-level risk, highlighting potential failure points and their impact. FTA provides a quantitative risk evaluation that complements the formal verification process, ensuring a thorough system reliability assessment.</div></div>","PeriodicalId":49561,"journal":{"name":"Science of Computer Programming","volume":"244 ","pages":"Article 103286"},"PeriodicalIF":1.5,"publicationDate":"2025-02-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143580311","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Trusta: Reasoning about assurance cases with formal methods and large language models","authors":"Zezhong Chen ,&nbsp;Yuxin Deng ,&nbsp;Wenjie Du","doi":"10.1016/j.scico.2025.103288","DOIUrl":"10.1016/j.scico.2025.103288","url":null,"abstract":"<div><div>Assurance cases can be used to argue for the safety of products in safety engineering. In safety-critical areas, the construction of assurance cases is indispensable. We introduce the Trustworthiness Derivation Tree Analyzer (Trusta), a tool designed to enhance the development and evaluation of assurance cases by integrating formal methods and large language models (LLMs). The tool incorporates a Prolog interpreter and solvers like Z3 and MONA to handle various constraint types, enhancing the precision and efficiency of assurance case assessment. Beyond traditional formal methods, Trusta harnesses the power of LLMs including ChatGPT-3.5, ChatGPT-4, and PaLM 2, assisting humans in the development of assurance cases and the writing of formal constraints. Our evaluation, through qualitative and quantitative analyses, shows Trusta's impact on improving assurance case quality and efficiency. Trusta enables junior engineers to reach the skill level of experienced safety experts, narrowing the expertise gap and greatly benefiting those with limited experience. Case studies, including automated guided vehicles (AGVs), demonstrate Trusta's effectiveness in identifying subtle issues and improving the overall trustworthiness of complex systems.</div></div>","PeriodicalId":49561,"journal":{"name":"Science of Computer Programming","volume":"244 ","pages":"Article 103288"},"PeriodicalIF":1.5,"publicationDate":"2025-02-28","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143529038","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"LayoutOptimizer: A layout rendering performance optimizer for Android application","authors":"Yue Wu ,&nbsp;Zhentao He ,&nbsp;Qingnan Wang ,&nbsp;Yihui Wang ,&nbsp;Huaxiao Liu","doi":"10.1016/j.scico.2025.103287","DOIUrl":"10.1016/j.scico.2025.103287","url":null,"abstract":"<div><div>The perceived delays experienced by Android application users can have a significant impact on their overall experience. Slow UI rendering is a major factor causing perceived delays. Poorly implemented UI layouts can have a considerable impact on rendering performance. To optimize the rendering performance of a layout, an important way is to solve its hierarchy issues. While there are layout performance analysis tools available, they lack effective solutions for fixing hierarchy issues, which limits their ability to assist developers in resolving such issues. In this paper, we propose a novel approach called LayoutOptimizer, which can automatically identify and solve two common hierarchy issues in Android layouts. The evaluation based on 31 layouts from real-world apps demonstrates that LayoutOptimizer can effectively fix the two common hierarchy issues while ensuring visual consistency.</div></div>","PeriodicalId":49561,"journal":{"name":"Science of Computer Programming","volume":"244 ","pages":"Article 103287"},"PeriodicalIF":1.5,"publicationDate":"2025-02-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143510919","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Path-guided conformance test case generation for models with data and time using symbolic execution techniques","authors":"Boutheina Bannour ,&nbsp;Arnault Lapitre ,&nbsp;Pascale Le Gall","doi":"10.1016/j.scico.2025.103285","DOIUrl":"10.1016/j.scico.2025.103285","url":null,"abstract":"<div><div>This paper presents an approach leveraging symbolic execution techniques to generate test cases from models mixing data and time. Our methodology focuses on symbolic paths, satisfying a trace-determinism property, which allows testing behaviors in the presence of uninitialized state variables. We construct tree-like test cases around these test purposes, with verdicts on their leaves, meticulously crafting verdict conditions from symbolic execution path conditions encoding temporal data-dependent constraints. Our test case generation is implemented within the symbolic execution platform Diversity. Through experiments, we provide metrics and quantify some aspects of the generated test cases, including the reachability of verdicts within observation time frames specified by the tester.</div></div>","PeriodicalId":49561,"journal":{"name":"Science of Computer Programming","volume":"243 ","pages":"Article 103285"},"PeriodicalIF":1.5,"publicationDate":"2025-02-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143480120","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"SBD: Securing safe rust automatically from unsafe rust","authors":"Shaowen Li,&nbsp;Hiroyuiki Sato","doi":"10.1016/j.scico.2025.103281","DOIUrl":"10.1016/j.scico.2025.103281","url":null,"abstract":"<div><div>System programming expects programmers to have fine control over available resources to ensure both the correctness and efficiency of the system. Programming languages designed for this type of task provide little abstraction of the underlying hardware. With greater power to interact directly with the machine comes greater responsibility for programmers to manage the hardware themselves to avoid any undefined behavior. C and C++ have been the long-standing de facto languages in this field as they offer both the programming experience of a modern language and the ability to manipulate low-level resources with the abstraction of pointers. However, this responsibility is demanding for programmers, leading to numerous bugs caused by improper resource management.</div><div>Rust is a rising system programming language aiming to combine both low-level resource manipulation and high-level resource management. The design philosophy of Rust is to make the compiler play a vital role in resource management. A set of static analysis unique to Rust are performed at compile time to ensure resources are handled correctly without runtime cost. Nevertheless, static analysis is inherently conservative and Rust addresses this by providing a feature called <em>unsafe Rust</em>, which is exempt from its strict static checks. Various unsafe operations, such as raw pointer dereferencing and foreign function calls, are only permitted within an unsafe code block. This is essential to make the language sufficiently expressive. Nonetheless, Rust's <em>unsafe block</em> only matters statically for type checking, without any runtime assurance. As a consequence, the effects of unsafe operations within an unsafe block can spread to the outside safe code and jeopardize its safety.</div><div>We present <em><u>S</u>afety <u>B</u>lock <u>D</u>ivision</em> (SBD), a completely automatic solution to isolate safe Rust from unsafe Rust. The fundamental design of SBD is its safety data-flow analysis performed on Rust intermediate representation (IR) to fully incorporate language features. This distinguishes SBD from previous works. Past designs primarily operate on LLVM IR and require manual efforts or external tools. SBD is entirely built into the Rust compiler, and thus no programmer involvement is required. We extensively evaluate SBD on popular Rust crates (libraries). Our experiments reveal that SBD incurs negligible binary size (0.31% increase on average) and runtime (a geometric mean of 7.52% increase across eight benchmarks) overhead. We also demonstrate that SBD is capable of protecting against real-world vulnerabilities.</div></div>","PeriodicalId":49561,"journal":{"name":"Science of Computer Programming","volume":"243 ","pages":"Article 103281"},"PeriodicalIF":1.5,"publicationDate":"2025-02-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143454135","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Extending the EB4EB framework with parameterised events","authors":"Peter Rivière,&nbsp;Neeraj Kumar Singh,&nbsp;Yamine Ait-Ameur,&nbsp;Guillaume Dupont","doi":"10.1016/j.scico.2025.103279","DOIUrl":"10.1016/j.scico.2025.103279","url":null,"abstract":"<div><div>EB4EB, standing for <em>Event-B for Event-B</em>, is a framework that supports the formalisation of Event-B models using first-order logic and set-theory, so that it becomes possible to manipulate them as first-class objects. This framework relies on an Event-B algebraic theory, which serves as a meta-theory formalising, explicitly, all of the features of an Event-B machine. States, events, invariants, variants, etc... are formalised through data-types and operators. When this meta-theory is instantiated, an Event-B model becomes a first-order logic and set-theoretic formula, described in an Event-B context.</div><div>Because it can handle machine elements as terms in formulas, the EB4EB framework enables the definition of new specific proof obligations and <em>analyses</em>. Such analyses may then be applied to any EB4EB machines in order to establish advanced properties, not natively present in the Event-B method, such as deadlock-freeness or liveness requirements. These analyses are <em>non-intrusive</em> since they do not require to alter the machine in order to be performed.</div><div>In the previous formalisation of the EB4EB framework only states and events were handled, limiting the expressive reasoning power of the framework. This paper presents an extension of the EB4EB framework to support parameterised events, an important feature of Event-B. This extension is not straightforward in EB4EB. Indeed, the typing system supported by Event-B theories is not rich enough to describe such extension in a constructive manner as for the other Event-B features formalised in EB4EB. The proposed solution, described in this paper, consists in defining an axiomatic formalisation of event parameters definitions. We also show that the proof obligations and model analyses we have defined scale to handle event parameters. The approach is illustrated on different case studies we have developed.</div></div>","PeriodicalId":49561,"journal":{"name":"Science of Computer Programming","volume":"243 ","pages":"Article 103279"},"PeriodicalIF":1.5,"publicationDate":"2025-02-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143480241","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Graph neural network-based long method and blob code smell detection","authors":"Minnan Zhang ,&nbsp;Jingdong Jia ,&nbsp;Luiz Fernando Capretz ,&nbsp;Xin Hou ,&nbsp;Huobin Tan","doi":"10.1016/j.scico.2025.103284","DOIUrl":"10.1016/j.scico.2025.103284","url":null,"abstract":"<div><div>The concept of code smell was first proposed in the late nineties, to refer to signals that code may need refactoring. While not necessarily affecting functionality, code smell can hinder understandability and future scalability of the program. As a result, the precise detection of code smell has become an important topic in coding research. However, current detection methods are limited by imbalanced and industrial-irrelevant datasets, a lack of sufficient structural and logical information on the code, and simple model architecture. Given these limitations, this paper utilized an industry-relevant and sufficient dataset and then developed a graph neural network to better detect code smell. First, we identified Long Method and Blob as our research subjects due to their frequent occurrence and impacts on the maintainability of software. We then designed modified fuzzy sampling with focalloss to address the issue of data imbalance. Second, to deal with the large volume of data, we proposed a global and local attention scoring mechanism to extract the key information from the code. Third, in order to design a graph neural network specifically for the abstract syntax tree of code, we combined Euclidean space and non-Euclidean space. Finally, we compared our method with other machine learning methods and deep learning methods. The results demonstrate that our method outperforms the other methods on Long Method and Blob, which indicates the effectiveness of our proposed method.</div></div>","PeriodicalId":49561,"journal":{"name":"Science of Computer Programming","volume":"243 ","pages":"Article 103284"},"PeriodicalIF":1.5,"publicationDate":"2025-02-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143488855","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Building augmented reality games with argDSL","authors":"Rubén Campos-López,&nbsp;Esther Guerra,&nbsp;Juan de Lara","doi":"10.1016/j.scico.2025.103271","DOIUrl":"10.1016/j.scico.2025.103271","url":null,"abstract":"<div><div>Augmented Reality (AR) has become popular. It does not require advanced technology, but only a mobile device with a camera to interact with virtual objects. However, developing AR applications – especially games – is time-consuming and requires in-depth knowledge of highly specialised technologies, and mathematical concepts related to the graphics and physics of the virtual objects.</div><div>To address this problem, we propose the domain-specific language <figure><img></figure>for creating AR games. It allows customising the game logic, and the virtual objects' domain, physics and representation. We provide an Eclipse editor to define AR games using the language, and an iOS client to run the games. Our experiments show the versatility of our proposal and the usability of the games.</div></div>","PeriodicalId":49561,"journal":{"name":"Science of Computer Programming","volume":"243 ","pages":"Article 103271"},"PeriodicalIF":1.5,"publicationDate":"2025-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143562614","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}