In recent years, GPUs have become the preferred accelerators for HPC and ML applications due to their parallelism and high memory bandwidth. While GPUs boost computation, inter-GPU communication can create scalability bottlenecks, especially as the number of GPUs per node and cluster grows. Traditionally, the CPU managed multi-GPU communication, but advancements in GPU-centric communication now challenge this CPU dominance by reducing its involvement, granting GPUs more autonomy in communication tasks, and addressing mismatches in multi-GPU communication and computation. This paper provides a landscape of GPU-centric communication, focusing on vendor mechanisms and user-level library supports. It aims to clarify the complexities and diverse options in this field, define the terminology, and categorize existing approaches within and across nodes. The paper discusses vendor-provided mechanisms for communication and memory management in multi-GPU execution and reviews major communication libraries, their benefits, challenges, and performance insights. Then, it explores key research paradigms, future outlooks, and open research questions. By extensively describing GPU-centric communication techniques across the software and hardware stacks, we provide researchers, programmers, engineers, and library designers insights on how to exploit multi-GPU systems at their best.

The rapid deployment of large language and vision models in real-world applications has intensified the need to address hallucinations—instances where models generate incorrect or incoherent outputs. These failures can spread misinformation and degrade workflows, causing financial and operational harm. Despite extensive research efforts, our understanding of hallucinations remains limited and fragmented. Without clear understanding, solutions risk addressing disparate symptoms rather than root causes, which undermines their effectiveness and generalisability during deployment. To address this, we first introduce a unified, multi-level framework to characterise both image and text hallucinations across broad applications, helping reduce conceptual fragmentation. Then, we trace their root causes to identifiable mechanisms within a model’s lifecycle in a task-modality interleaved manner, fostering a deeper and more holistic understanding. Our investigations reveal hallucinations as predictable consequences of underlying distributions and biases. By enhancing our understanding of hallucinations, this survey lays the groundwork for more effective solutions to hallucinations in generative AI systems.

Large language models (LLMs) have advanced rapidly, emerging as versatile tools across fields thanks to their exceptional language understanding, generation, and reasoning capabilities. However, performing LLM inference at the network edge remains challenging due to their large memory and compute demands. This survey outlines the challenges specific to LLM edge inference and provides a comprehensive overview of recent progress, covering system architectures, model optimization and deployment, and resource management and scheduling. By synthesizing state-of-the-art techniques and mapping future directions, this survey aims to unlock the potential of LLMs in resource-constrained edge environments.

Cyber-Physical Systems (CPS), in which computation and physical processes converge, have found application in many domains and will only gain further importance. Faced with such developments, recent research emphasized the need to not lose sight of affected humans. Two important notions to that respect are human-centric and socio-technical CPS. Moving towards such a vision poses challenges concerning how CPS design should best be done. Therefore, we conducted a systematic review to identify and qualitatively evaluate CPS design support showing human-centric and/or socio-technical characteristics, focusing on its embedding into design processes. We first conceptualized the two terms, trying to account for conceptual ambiguity, and then characterized each identified work. We recognized the central role of interactions between human and machine components for both paradigms, with design activities targeting different levels of abstraction. Positive reports on active involvement of affected humans indicated future research potential. Investigated concepts and ideas relating to CPS architectures that enable design support showed the need to properly address inherent CPS complexity. We also found a need to focus validation activities on achieved outcomes for targeted design support user groups as well as people affected by the CPS.

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.

This systematic literature review explores the landscape of risks and risk management techniques in cloud outsourcing, with a focus on assisting enterprise cloud consumers in understanding and mitigating both technical and non-technical risks, despite having limited control over the infrastructures. From a comprehensive analysis of 55 academic articles, spanning the period from January 2013 to September 2022, we identify and characterize risks using established frameworks from ENISA and [20]. Using ISO31000 and the classification proposed by [4], we also summarize and characterize 23 main strategies in risk management techniques feasible for cloud consumers, including technical and non-technical measures. We observe a significant emphasis on technical risks in the literature, while non-technical risks, including legal, organizational, and policy aspects, are relatively underrepresented. Threats to data confidentiality dominate the technical risks and mostly originate from shared infrastructure issues. However, non-technical issues, such as vendor lock-in, also pose catastrophic risks the continuity and business operations of the cloud consumers. We also observe that encryption still plays a key role in the existing techniques, next to other techniques such as auditing, risk-aware software development, and assessments of third parties.

Video generation aims to produce temporally coherent sequences of visual frames, representing a pivotal advancement in Artificial Intelligence Generated Content (AIGC). Compared to static image generation, video generation poses unique challenges: it demands not only high-quality individual frames but also strong temporal coherence to ensure consistency throughout the spatiotemporal sequence. Although research addressing spatiotemporal consistency in video generation has increased in recent years, systematic reviews focusing on this core issue remain relatively scarce. To fill this gap, this paper views the video generation task as a sequential sampling process from a high-dimensional spatiotemporal distribution, and further discusses spatiotemporal consistency. We provide a systematic review of the latest advancements in the field. The content spans multiple dimensions including generation models, feature representations, generation frameworks, post-processing techniques, training strategies, benchmarks and evaluation metrics, with a particular focus on the mechanisms and effectiveness of various methods in maintaining spatiotemporal consistency. Finally, this paper explores future research directions and potential challenges in this field, aiming to provide valuable insights for advancing video generation technology. The project link is https://github.com/Yin-Z-Y/A-Survey-Spatiotemporal-Consistency-in-Video-Generation

Security engineering—from creating security requirements to the implementation of security features, such as cryptography or authentification—is often supported by domain-specific languages (DSLs). While many security DSLs have been presented, a lack of overview and empirical data about these DSLs, such as which security aspects are addressed and when, hinders their effective use and further research. This systematic literature review examines 120 security DSLs regarding their security aspects and goals addressed, their language-specific characteristics, their integration into the software development lifecycle, and their evaluation. We observe a focus on individual development phases and a high degree of fragmentation, which leads to opportunities for integration. The research community also needs to improve the usability and evaluation of security DSLs.

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.

Traditional sociological research often relies on human participation, which, though effective, is expensive, challenging to scale, and with ethical concerns. Recent advancements in large language models (LLMs) highlight their potential to simulate human behavior, enabling the replication of individual responses and facilitating studies on many interdisciplinary studies. In this paper, we conduct a comprehensive survey of this field, illustrating the recent progress in simulation driven by LLM-empowered agents. We categorize the simulations into three types: (1) Individual Simulation , which mimics specific individuals or demographic groups; (2) Scenario Simulation , where multiple agents collaborate to achieve goals within specific contexts; and (3) Society Simulation , which models interactions within agent societies to reflect the complexity and variety of real-world dynamics. These simulations follow a progression, ranging from detailed individual modeling to large-scale societal phenomena. We provide a detailed discussion of each simulation type, including the architecture or key components of the simulation, the classification of objectives or scenarios, and the evaluation method. Afterward, we summarize commonly used datasets, benchmarks and tools. Finally, we discuss the risks and challenges across these three types of simulation. A repository for the related sources is at https://github.com/FudanDISC/SocialAgent.