Blockchain technologies have experienced rapid adoption across various sectors, including supply chain management, decentralized finance and cross-border payments. With this growth, however, the complexity and security risks of maintaining blockchain integrity and functionality have increased. Addressing these challenges requires a systematic and rigorous organization of knowledge in blockchain security. This paper presents a Systematization of Knowledge (SoK) study based on a structured survey of academic literature, industry reports and real-world case studies. The authors classify vulnerabilities into three layers: system-level, smart contract-level and application-level, analyzing their root causes, real-world prevalence and mitigation tradeoffs. The taxonomy encompasses blockchain-specific threats (e.g. gas-based DoS attacks, MEV) as well as vulnerabilities inherited from distributed systems and software (e.g. Sybil attacks, access control failures). The authors critically evaluate detection and mitigation techniques, including static and dynamic analysis, fuzzing, symbolic execution and formal verification, assessing their precision, recall, scalability and inherent limitations. The authors further review state-of-the-art auditing tools in terms of methodology, adoption and shortcomings. Finally, the authors discuss advanced approaches such as hybrid frameworks that combine AI with program analysis, SMT solvers, and zero-knowledge proofs, outlining how these can address current gaps in scalability, interpretability and runtime verification. Overall, this study systematizes the security landscape of blockchain technologies, synthesizes the limitations of current approaches, and identifies technically actionable future research directions toward building more robust and resilient blockchain systems.
The notion of Zero Trust Architecture (ZTA) has been introduced as a fine-grained defense approach. It assumes that no entities outside and inside the protected system can be trusted and, therefore, requires articulated and high-coverage deployment of security controls. However, ZTA is a complex notion that does not have a single design solution; rather, it consists of numerous interconnected concepts and processes that need to be assessed prior to deciding on a solution. In this monograph, we cover the principles and architectural foundations of ZTA, basically following the guidelines by NIST, and provide a detailed analysis of ZT architectures proposed by research and industry. The monograph also describes an approach for the automatic generation of ZT policies based on application communication requirements, network topology, and organizational information. This approach was designed to meet a critical need of ZTA, that is, the generation and implementation of a large number of fine-grained policies. Finally, the monograph discusses several research directions, including the incorporation of threat intelligence into ZT networks and the use of large language models (LLMs).
Blockchains have become an integrated part of our finance infrastructures. Being monetary yet fully automated, blockchains and their applications are unanimously deemed impracticable before undergoing necessary verification. This monograph reviews the previous attempts at verifying two fundamental properties of blockchains: correctness (where flaws lead to unintentional damages) and security (where vulnerabilities incur attacks and losses). First, it summarizes and categorizes the correctness and security flaws encountered by real-world blockchains. Second, it systematizes the development of formal verification to address the flaws in blockchains, covering the aspects of models, specifications, and techniques. Third, it unveils the progress of security analysis for mitigating the flaws, unveiling the analysis principles being followed, the flaw oracles being devised, and the detection methods being used. Finally, it summarizes the challenges remaining to be addressed, followed by our vision of the trend in the near future. Throughout this monograph, we anticipate shedding light on future blockchain verification advances, especially in expanding its applicability, making specification generation easier, and discovering previously unknown vulnerabilities. By identifying gaps such as missing tools for infrastructure-level components and the difficulty of writing formal specifications, this work aims to motivate the development of more automated, intelligent, and practical verification frameworks.
The success of machine learning algorithms relies not only on achieving good performance but also on ensuring trustworthiness across diverse applications and scenarios. Trustworthy machine learning seeks to handle critical problems in addressing the issues of robustness, privacy, security, reliability, and other desirable properties. The broad research area has achieved remarkable advancement and brings various emerging topics along with the progress. We present this survey to provide a systematic overview of the research problems under trustworthy machine learning covering the perspectives from data to model. Starting with fundamental data-centric learning, the survey reviews learning with noisy data, long-tailed distribution, out-of-distribution data, and adversarial examples to achieve robustness. Delving into private and secured learning, the survey elaborates on core methodologies differential privacy, different attacking threats, and learning paradigms, to realize privacy protection and enhance security. Finally, it introduces several trendy issues related to the foundation models, including jailbreak prompts, watermarking, and hallucination, as well as causal learning and reasoning. The survey integrates commonly isolated research problems in a unified manner, which provides general problem setups, detailed sub-directions, and further discussion on its challenges or future developments. We hope the comprehensive investigation presented in this survey can serve as a clear introduction for the problemevolution from data to models and also bring new insight for developing trustworthy machine learning.
In recent years, the integration of Large Language Models (LLMs) and Recommender Systems (RS) has revolutionized the way personalized and intelligent user experiences are delivered. This survey provides an extensive review of critical challenges, current landscape, and future directions in the collaboration between LLM-based AI agents (LLM Agent) and recommender systems. We begin with an introduction to the foundational knowledge, exploring the components of LLM agents and the applications of LLMs in recommender systems. The survey then delves into the symbiotic relationship between LLM agents and recommender systems, illustrating how LLM agents enhance recommender systems and how recommender systems support better LLM agents. Specifically, we discuss the overall architectures for designing LLM agents for recommendation, encompassing profile, memory, planning, and action components, along with multi-agent collaboration. Conversely, we investigate how recommender systems contribute to LLM agents, focusing on areas such as memory recommendation, plan recommendation, tool recommendation, agent recommendation, and personalized LLMs and LLM agents. Furthermore, a critical evaluation of trustworthy AI agents and recommender systems follows, addressing key issues of safety, explainability, fairness, and privacy. Finally, we propose potential future research directions, highlighting emerging trends and opportunities in the intersection of AI agents and recommender systems. This survey concludes by summarizing the key insights of current research and outlining promising avenues for future exploration in this rapidly evolving field. A curated collection of relevant papers for this survey is available in the GitHub repository: https://github.com/agiresearch/AgentRecSys.
