Fully Homomorphic Encryption (FHE) enables arbitrary computations on encrypted data but lacks mechanisms to ensure the integrity of those computations. In particular, verifying that algorithm inputs are correct or that the intended algorithm was indeed executed remains an open challenge. This article addresses the issue by making two key contributions. First, we perform a comprehensive analysis of integrity faults in FHE, culminating in the definition of verifiable FHE as a novel concept to tackle these concerns. Second, we present a systematic review of existing approaches aimed at providing verifiable FHE, assessing their strengths and weaknesses, as well as their applicability in practical settings. Our findings reveal that, despite promising advances, significant gaps persist in both the theoretical foundations and the practical deployment of verifiable FHE. We conclude by outlining future research directions necessary to achieve verifiable FHE computations.

Cloud-native computing is a software development approach that uses cloud computing to build and run scalable applications by leveraging best practices and technologies in the field such as DevOps, general unified architectures (e.g., based on serverless functions, containers, container orchestrators, sidecar, service meshes, microservices), agile software development and so on. It is also used for application development environments that enable an abstraction layer on top of non-cloud provider services. Cloud-native is expected to increase the efficiency of development, operation, and verification of 5G and beyond network functionalities along with automated service upgrades and deployments, and ultimately reduce vendor lock-in so that applications can be easily migrated to other environments seemlessly. In this paper, we provide a state-of-the-art overview of existing cloud-native approaches for telcos, their key features, the role of open-source and Linux Foundation projects, and all associated standardization efforts. After presenting key findings of the recent research studies and related implications under “Lessons Learned”, we highlight the potential challenges, current issues, and future directions of cloud-native computing for telecommunication networks.

The rapid integration of AI systems into high-stakes domains has revealed persistent issues of user distrust, algorithmic aversion, and over-reliance, highlighting the need for decision-making frameworks in which humans and machines synergistically collaborate towards the solution of the task. Hybrid Decision-Making Systems (HDMS) have emerged as a paradigm where humans and AI jointly contribute to the same task, leveraging and integrating human strengths like domain expertise, contextual understanding and flexible reasoning, alongside machines’ computational power. This survey offers a structured overview of learning paradigms for HDMS, with a particular focus on uncertainty-driven abstention mechanisms, which determine when an AI system should act autonomously or when it should call for human intervention. We formalise and compare algorithmic approaches that embed machine learning models with the capacity to “know what they don’t know”, analysing how abstention policies and system architectures integrate human expertise into the decision pipeline. Beyond abstention, we examine frameworks that support direct human–machine interaction during and after the learning process, outlining emerging approaches that foster bidirectional collaboration between humans and AI. Building on this analysis, we propose a taxonomy of three learning paradigms characterising progressively tighter human–machine integration.

Software defect datasets, which are collections of software bugs, are essential resources to facilitate empirical research and enable standardized benchmarking for a wide range of software engineering techniques, including emerging areas like agentic AI-based software development. Over the years, numerous software defect datasets have been developed, providing rich resources for the community, yet making it increasingly difficult to navigate the landscape. This article provides a comprehensive survey of 151 software defect datasets, covering their scope, construction, availability, usability, and practical uses. We also suggest potential opportunities for future research based on our findings, such as addressing underrepresented kinds of defects. A complete catalog of all surveyed software defect datasets is available at https://defect-datasets.github.io/ .

With the rapid development of artificial intelligence, music generation has evolved from single-modal to cross-modal approaches and is gradually moving toward multi-modal fusion. This survey systematically reviews this developmental trajectory. The discussion begins with the representation methods for key modalities, including audio, symbolic, text, and visual data. Music generation techniques are then organized across single-modal, cross-modal, and multi-modal settings. In addition, key datasets and evaluation methodologies relevant to these tasks are compiled. Finally, the survey discusses core challenges in the field, including modal fusion, data scarcity, and evaluation frameworks, and outlines potential directions for future research.

In order to gain an overview of the state of research using machine learning for applications in cybersecurity, we have carried out a quantitative analysis of published works using the Scopus database. In total, 81, 082 articles in this area were scraped from Scopus, from which we created an article-citation graph with articles as nodes and 547, 998 citations among them as edges. We used the Louvain method for community-finding to organize the articles into research areas based on citation patterns, resulting in 12 identified research areas: intrusion detection: classical machine learning, intrusion detection: deep learning, privacy-preserving machine learning, malware detection, biometrics, adversarial machine learning, steganalysis, neural cryptography, phishing, traffic classification, software vulnerabilities, and smart grid. The state of the whole field and of each research area is discussed in detail, including highly cited and seminal works in each area and the relative size of each area compared to each other. Compared to previous literature surveys on this topic this review is not limited to specific subtopics and organizes the field by grouping works based on research topics, giving a more comprehensive overview of the field and the areas it encompasses.

Knowledge and information in the biomedical domain are predominantly represented in text format and available in diverse sources, such as articles, patents, electronic health records, and social media. Automated approaches for organising these resources face significant challenges when understanding natural language. Knowledge organisation systems serve as a crucial standard for enhancing information retrieval and sharing among humans and automated approaches. Named entity linking plays a crucial role in bridging the gap between text and knowledge organisation systems: it maps relevant entities described in text, such as diseases, chemicals, and genes, to unambiguous entries in target knowledge organisation systems that accurately describe their meaning. In this review, we analysed 102 articles published between 2013 and 2024 related to named entity linking, which provides an overview of the landscape of knowledge organisation systems in the biomedical and clinical domains and of the evolution of this task over the past decade. The findings highlight key limitations of existing approaches and identify promising directions for future research.

Given a set S of n keys, a perfect hash function for S maps the keys in S to the first m ≥ n integers without collisions. It may return an arbitrary result for any key not in S and is called minimal if m = n . The most important parameters are its space consumption, construction time, and query time. Years of research now enable modern perfect hash functions to be extremely fast to query, very space-efficient, and scale to billions of keys. Different approaches give different trade-offs between these aspects. For example, the smallest constructions get within 0.1% of the space lower bound of log 2 e bits per key. Others are particularly fast to query, requiring only one memory access. Perfect hashing has many applications, for example to avoid collision resolution in static hash tables, and is used in databases, bioinformatics, and stringology. Since the last comprehensive survey in 1997, significant progress has been made. This survey covers the latest developments and provides a starting point for getting familiar with the topic. Additionally, our extensive experimental evaluation can serve as a guide to select a perfect hash function for use in applications.

The dynamic nature of modern network settings presents challenges for traditional base stations, leading to inefficient resource utilisation. These issues are exacerbated by the unpredictability of network scenarios, such as unexpected hardware problems or sudden increases in user demand. More adaptive and intelligent systems that can handle these issues are becoming increasingly critical and essential. Open-RAN is an intelligent and adaptable mobile network architecture featuring learning-based capabilities and a software layer for adaptation that can incorporate AI and ML to optimise communication networks and improve resource efficiency. Hence, this article aims to systematically review intelligence for Open-RAN through the adoption of the RAN Intelligent Controller (RIC). This comprehensive review also includes (1) real-world use cases for various industries, (2) the potential use cases for various intelligent models in terms of 5G core, RAN and air interface, (3) the advantages and challenges of AI incorporation within Open-RAN, and (4) the societal, environmental and economic impact of Open-RAN intelligence. This article outlines the significance of intelligence in the context of Open-RAN and highlights potential use cases, examining their impact and presenting challenges for further exploration by researchers in the field.

This survey presents a comprehensive analysis of Adversarial Machine Learning (AML) as a critical approach for enhancing the security of Internet of Things (IoT) ecosystems, with a focus on its integration into Intrusion Detection Systems (IDSs). It synthesizes existing research, identifies current limitations, and outlines future directions for effective AML deployment. The survey explores emerging technologies to improve IDS robustness, adaptability, and threat anticipation capabilities. It further addresses the challenges of implementing AML in resource-constrained IoT environments. A structured, multi-phase attack framework inspired by the MITRE ATT&CK model is proposed to illustrate the evolving lifecycle of adversarial threats. Ethical and regulatory considerations are examined, emphasizing responsible deployment and accountability. The cross-domain applicability of AML is discussed across various environments, including smart grids, autonomous systems, cloud, and edge computing. Real-world use cases are analyzed, with attention to adversarial datasets and evaluation practices. Additionally, the role of Human-in-the-Loop (HITL) systems is highlighted for enhancing interpretability and trust in AML-powered IDSs. This survey serves as a valuable resource for researchers and practitioners aiming at advancing AML in securing IoT ecosystems with innovative, ethical, and practical defense strategies.