ACM Computing Surveys最新文献

{"title":"A Survey of Deep Learning Based Software Refactoring","authors":"Bridget Nyirongo, Yanjie Jiang, He Jiang, Hui Liu","doi":"10.1145/3849693","DOIUrl":"https://doi.org/10.1145/3849693","url":null,"abstract":"Refactoring is a critical activity in software engineering used to improve software quality and maintainability. Increasingly, deep learning (DL) techniques are being applied to software refactoring to overcome the limitations of traditional, manually designed heuristics. However, there is a lack of comprehensive surveys and structured taxonomies characterizing these deep learning-based approaches. To fill this gap, we present a survey that categorizes deep learning-based software refactoring into five core tasks: code smell detection, refactoring recommendation, refactoring implementation, refactoring mining, and refactoring impact analysis. Our analysis reveals an imbalance in research distribution. The field is concentrated on the early decision-making stages, with code smell detection ( <jats:inline-formula content-type=\"math/tex\"> <jats:tex-math notation=\"TeX\" version=\"MathJaX\">(43.84% )</jats:tex-math> </jats:inline-formula> ) and refactoring recommendation ( <jats:inline-formula content-type=\"math/tex\"> <jats:tex-math notation=\"TeX\" version=\"MathJaX\">(32.88% )</jats:tex-math> </jats:inline-formula> ) comprising more than three-quarters of the literature. Conversely, refactoring implementation ( <jats:inline-formula content-type=\"math/tex\"> <jats:tex-math notation=\"TeX\" version=\"MathJaX\">(12.33% )</jats:tex-math> </jats:inline-formula> ) and mining ( <jats:inline-formula content-type=\"math/tex\"> <jats:tex-math notation=\"TeX\" version=\"MathJaX\">(8.22% )</jats:tex-math> </jats:inline-formula> ) show emergent tracking, while refactoring impact analysis remains under-explored ( <jats:inline-formula content-type=\"math/tex\"> <jats:tex-math notation=\"TeX\" version=\"MathJaX\">(2.74% )</jats:tex-math> </jats:inline-formula> ). Beyond this taxonomy, the survey highlights critical gaps, including a notable absence of research on quality assurance for DL-based refactoring and minimal attention given to class- and variable-level operations. Finally, we map out open challenges and future directions to guide researchers toward building more comprehensive, end-to-end automated refactoring pipelines.","PeriodicalId":50926,"journal":{"name":"ACM Computing Surveys","volume":"16 1","pages":""},"PeriodicalIF":16.6,"publicationDate":"2026-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149086571","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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":"A Survey of Robotic Navigation and Manipulation with Physics Simulators in the Era of Embodied AI","authors":"Lik Hang Kenny Wong, Xueyang Kang, Kaixin Bai, Jianwei Zhang","doi":"10.1145/3856802","DOIUrl":"https://doi.org/10.1145/3856802","url":null,"abstract":"Navigation and manipulation are core capabilities in Embodied AI, but training agents to perform them directly in the real world is costly, time-consuming, and unsafe. Therefore, sim-to-real transfer has emerged as a key approach, yet the sim-to-real gap persists. This survey examines how physics simulators address this gap by analyzing properties that have received limited attention in prior surveys. We also analyze their features for navigation and manipulation tasks, as well as their hardware requirements. Additionally, we offer a resource with benchmark datasets, metrics, simulation platforms, and methods to help researchers select suitable tools while accounting for hardware constraints.","PeriodicalId":50926,"journal":{"name":"ACM Computing Surveys","volume":"148 1","pages":""},"PeriodicalIF":16.6,"publicationDate":"2026-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149086569","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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":"Complementary-Label Learning: A Survey of Methods, an Open-Source Toolkit, and Fair Benchmarking Results","authors":"Ha Mai Tan, Nai-Xuan Ye, Wei Wang, Gang Niu, Takashi Ishida, Masashi Sugiyama, Hsuan-Tien Lin","doi":"10.1145/3847508","DOIUrl":"https://doi.org/10.1145/3847508","url":null,"abstract":"<jats:italic toggle=\"yes\">Complementary-label learning</jats:italic> (CLL) is a weakly supervised learning paradigm designed for multiclass classification tasks. In CLL, algorithms are provided with complementary labels, which indicate the classes an instance does not belong to, rather than specifying its true class. With the growing research interest in CLL and the unique challenges it presents, particularly in handling incomplete and indirect supervision, a comprehensive and systematic overview of existing methodologies, results, and research directions has become increasingly essential. In this survey, we provide the first systematic synthesis of recent advances in CLL and offer three major contributions. First, we establish a structured taxonomy that clarifies foundational concepts, distinctive algorithmic strategies, key contributions, and their inherent limitations. Second, to support reproducibility and accelerate future research, we introduce a publicly available <jats:italic toggle=\"yes\">open-source CLL toolkit</jats:italic> that unifies key algorithm implementations and offers ready-to-use benchmarking pipelines. Third, we conduct an <jats:italic toggle=\"yes\">extensive benchmarking</jats:italic> study covering a wide spectrum of CLL algorithms, using standardized evaluation protocols across both synthetic, human-annotated, and VLM-annotated complementary-label datasets. This provides the most complete performance comparison currently available. Finally, we discuss current open research challenges and identify promising avenues for future investigation, emphasizing improvements to the practical applicability and robustness of CLL in real-world scenarios. Overall, this survey provides an up-to-date, in-depth resource on the evolving CLL landscape.","PeriodicalId":50926,"journal":{"name":"ACM Computing Surveys","volume":"34 778 1","pages":""},"PeriodicalIF":16.6,"publicationDate":"2026-10-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149086572","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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":"The Impact of Explanation Characteristics on the Effects of Recommender Systems Explanations","authors":"Kathrin Wardatzky, Oana Inel, Luca Rossetto, Athina Kyriakou, Abraham Bernstein","doi":"10.1145/3856207","DOIUrl":"https://doi.org/10.1145/3856207","url":null,"abstract":"There seems to be a consensus that adding explanations to AI systems, such as recommender systems, can have positive effects, such as increasing a user’s trust, the transparency of the system, or the efficiency of a decision-making process. It is to this date unclear, though, how an explanation needs to be designed to evoke these effects. To address this gap, we conducted a systematic literature survey of 180 articles. The corpus was created by first querying six literature databases with predefined keyword searches and sorting the retrieved articles by applying inclusion and exclusion criteria. We then extracted and categorized the explanation characteristics, such as their modalities and content, as well as the effects measured in user evaluations. Finally, we combine both dimensions to analyze the impact of the characteristics on the effects. The findings suggest that it is not possible to make general statements about the effects of explanations. Instead, the measured effects vary in different contexts, and the results depend on whether the evaluated explanation was compared to another explanation or to a system without explanation. Our survey identifies and highlights research gaps and promising areas for future work, such as the effects of non-local explanations on the user.","PeriodicalId":50926,"journal":{"name":"ACM Computing Surveys","volume":"2 1","pages":""},"PeriodicalIF":16.6,"publicationDate":"2026-09-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149077696","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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":"A Systematic Review of Digital Twins and Security for Cyber-Physical Systems","authors":"Arda Goknil, Ivan Kurtev, Leonard Tudorache, Adela Nedisan Videsjorden, Rustem Dautov, Hui Song, Dumitru Roman, Mihai Gheorghe, Ahmet Soylu","doi":"10.1145/3856206","DOIUrl":"https://doi.org/10.1145/3856206","url":null,"abstract":"The rapid growth of Cyber-Physical Systems (CPS) has increased interest in digital twins (DTs) as virtual representations of physical systems to improve efficiency, monitoring, and security. However, as DTs become integral components of CPS, ensuring their own security also becomes critical. In this paper, we present a systematic literature review (SLR) on DTs and security for CPS. Following a standard SLR methodology, we defined three research questions, performed a structured search in four publication databases complemented by forward snowballing, applied inclusion and exclusion criteria, and extracted and synthesized data from the selected studies. The review examines both how DTs are used to enhance CPS security and how DTs themselves are secured within CPS. Our main results are: (i) a classification of CPS security challenges, threats, and vulnerabilities addressed using DTs; (ii) a mapping of security mechanisms, application domains, and emerging trends in using DTs for CPS security; (iii) a classification of DT security challenges, threats, and vulnerabilities across architectural, operational, data, and trust dimensions; (iv) a synthesis of state-of-the-art approaches for securing DTs; and (v) an analysis of open research problems and future directions for advancing both CPS security through DTs and the security of DTs themselves.","PeriodicalId":50926,"journal":{"name":"ACM Computing Surveys","volume":"17 1","pages":""},"PeriodicalIF":16.6,"publicationDate":"2026-09-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149077699","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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":"Brain-Computer Interface Software Ecology: Open-Source Tools, Cross-Domain Applications, and Future Challenges","authors":"Zhenzhen Wu, Tzyy-Ping Jung, Wenlong Jiao, Xueguang Xie, Xianglong Wan, Dingna Duan, Tiange Liu, Suhan Cui, Dong Wen","doi":"10.1145/3833873","DOIUrl":"https://doi.org/10.1145/3833873","url":null,"abstract":"Open-source software platforms have revolutionized brain-computer interface (BCI) development. This comprehensive review examines the ecosystem for non-invasive EEG-based BCIs through four dimensions: technological foundations, tool taxonomy, scenario adaptation, and future evolution. The study demonstrates how open-source software enables complete BCI workflows from signal acquisition to real-time interaction. A systematic framework categorizes tools for applications in healthcare, education, and entertainment. Technical challenges and breakthrough opportunities in the ecosystem are identified. The study offers developers methodologies to select open-source BCI tools. Collectively, these insights aim to accelerate the transition of BCI technologies from research settings to real-world deployment.","PeriodicalId":50926,"journal":{"name":"ACM Computing Surveys","volume":"22 1","pages":""},"PeriodicalIF":16.6,"publicationDate":"2026-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149027581","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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":"Publish/Subscribe for Edge Intelligence: Systematic Review and Future Prospects","authors":"Alaa Saleh, Sasu Tarkoma, Susanna Pirttikangas, Lauri Lovén","doi":"10.1145/3849800","DOIUrl":"https://doi.org/10.1145/3849800","url":null,"abstract":"The exponential growth of IoT has led to billions of devices generating massive data streams. However, the resource-limited nature of IoT devices, coupled with the rise of interactive applications, requires efficient processing across the Device-Edge-Cloud continuum. The publish/subscribe (pub/sub) model has emerged as a preferred paradigm due to its scalability and flexibility. This review critically analyzes pub/sub systems, examining their architectures, security mechanisms, protocols, and real-world implementations. Unlike previous surveys, we explore their role in EdgeOps and real-time applications, identifying key limitations and opportunities for enhancement. Furthermore, we highlight underexplored challenges, such as adaptive message routing, and discuss how AI-driven pub/sub frameworks could improve system autonomy and responsiveness. This survey provides a roadmap for advancing pub/sub mechanisms to meet the demands of interactive distributed computing environments. Additionally, we propose the integration of agentic AI to enhance adaptability and autonomy. We further identify several key challenges that have the potential to shape the future of the pub/sub model, and we present these as important open issues for further exploration.","PeriodicalId":50926,"journal":{"name":"ACM Computing Surveys","volume":"479 1","pages":""},"PeriodicalIF":16.6,"publicationDate":"2026-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149027578","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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":"Representation Control for Large Language Models: Survey and Research Challenges","authors":"Lukasz Bartoszcze, Sarthak Munshi, Bryan Sukidi, Jennifer Yen, Zejia Yang, David Williams-King, Linh Le, Carsten Maple","doi":"10.1145/3846173","DOIUrl":"https://doi.org/10.1145/3846173","url":null,"abstract":"Large language models (LLMs) are capable of completing a variety of tasks, but remain unpredictable and intractable. Representation Control (RepControl) seeks to resolve this problem through targeted interventions that modify high-level representations of concepts such as honesty, harmfulness or power-seeking. We formalize the goals and methods of RepControl to present a cohesive picture of work in this emerging field, focusing on techniques that steer model behavior by manipulating internal activations at inference time. We compare these control methods with alternative approaches, such as prompt-engineering and fine-tuning. We outline challenges such as performance degradation, computational overhead, and limitations in steering precision. We present a clear agenda for future research to build more steerable, personalized, and reliable LLMs through advances in RepControl techniques.","PeriodicalId":50926,"journal":{"name":"ACM Computing Surveys","volume":"80 1","pages":""},"PeriodicalIF":16.6,"publicationDate":"2026-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149027580","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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":"Frameworks, Modeling and Simulations of Misinformation and Disinformation: A Systematic Literature Review","authors":"Alejandro Buitrago López, Javier Pastor-Galindo, José A. Ruipérez-Valiente","doi":"10.1145/3849797","DOIUrl":"https://doi.org/10.1145/3849797","url":null,"abstract":"Misinformation and disinformation pose a significant challenge in today’s digital landscape. To examine how these phenomena are conceptualized and evaluated, this systematic literature review analyzes frameworks, models, and simulations used to represent mis/disinformation processes and dynamics. It covers 115 studies published through 2024 and retrieved from five bibliographic sources. Following the PRISMA methodology, the studies were analyzed to identify (1) the terminology and definitions of mis/disinformation, (2) the methods used to represent mis/disinformation, (3) the primary purpose beyond modeling and simulating mis/disinformation, (4) the application contexts in which mis/disinformation is studied, and (5) the strategies used to validate these approaches for understanding mis/disinformation. The findings show a broadly consistent definition of misinformation and disinformation across studies, with intent as the key distinguishing factor. Among the reviewed studies, social frameworks and epidemiological models are the most common approaches. Belief updating and mis/disinformation diffusion are the most frequent simulation types. The approaches are mainly used to conceptualize mis/disinformation and evaluate its effects or related countermeasures. Health and politics are the most common application contexts. Among validated approaches, comparison with real-world data is the most frequently reported strategy. Finally, this paper identifies current trends and open challenges and provides recommendations for future work.","PeriodicalId":50926,"journal":{"name":"ACM Computing Surveys","volume":"101 1","pages":""},"PeriodicalIF":16.6,"publicationDate":"2026-09-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149027579","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"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":"Multi-Task Deep Recommender Systems: A Survey","authors":"Yuhao Wang, Ha Tsz Lam, Yi Wong, Ziru Liu, Wanyu Wang, Maolin Wang, Yichao Wang, Bo Chen, Huifeng Guo, Ruiming Tang, Xiangyu Zhao","doi":"10.1145/3846376","DOIUrl":"https://doi.org/10.1145/3846376","url":null,"abstract":"Multi-task learning (MTL) aims at learning related tasks in a unified model to achieve mutual improvement among tasks considering their shared knowledge. It is an important topic in recommendation due to the demand for multi-task prediction considering performance and efficiency. Although MTL has been well studied and developed, there is still a lack of systematic review in the recommendation community. To fill the gap, we provide a comprehensive review of existing multi-task deep recommender systems (MTDRS) in this survey. To be specific, the problem definition of MTDRS is first given, and it is compared with other related areas. Next, the development of MTDRS is depicted and the taxonomy is introduced from the task relation and methodology aspects. Specifically, the task relation is categorized into parallel, cascaded, and auxiliary with main, while the methodology is grouped into parameter sharing, optimization, and training mechanism. The survey concludes by summarizing the application and public datasets of MTDRS and highlighting the challenges and future directions of the field.","PeriodicalId":50926,"journal":{"name":"ACM Computing Surveys","volume":"16 1","pages":""},"PeriodicalIF":16.6,"publicationDate":"2026-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"149020035","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}