Yingxuan Yang, Qiuying Peng, Jun Dan Wang, Ying Wen · 5 authors
Recent advances in large language models (LLMs) have enabled the development of LLM agents-autonomous systems capable of perceiving their environment, reasoning about tasks, and taking actions using external tools. While existing LLM-based Multi-Agent Systems (LaMAS) have shown promising results, they are predominantly centralized, operating within specific tasks or scenarios. These centralized designs simplify coordination but are fundamentally constrained by the limited data and knowledge available within a single entity. As LLM agents see broader deployment, the complexity of tasks increasingly requires collaboration across multiple organizations and data domains. Since organizations cannot and will not fully share their proprietary data, the next frontier of artificial intelligence lies in collective intelligence through decentralized LLM-based Multi-Agent Systems (LaMAS), where LLM agents, each accessing proprietary knowledge and tools, collaborate to solve complex tasks. This paradigm is becoming not just possible but necessary with the growing adoption of LLM agents across diverse organizations. This paper explores the transformative potential of decentralized LaMAS. In decentralized settings, two key issues arise: (1) privacy-preserving mechanisms that enable meaningful collaboration while safeguarding proprietary data and knowledge, and (2) monetization and credit attribution mechanisms that incentivize continuous improvement of agent capabilities and ensure fair value distribution among participants. Our analysis reveals that addressing these challenges can unlock a new paradigm of artificial collective intelligence that overcomes the limitations. This work contributes to decentralized AI by proposing a practical framework for mechanism design that advances both technological innovation and economic sustainability in decentralized LLM Agent networks.
This work explores the integration of Relationalmodel Multi-Agent System (RMAS) with Zero-Knowledge Proofs (ZKP) to enable verifiable computations in distributed systems. RMAS, based on relational algebra, simplifies multi-agent computations using active databases. By combining RMAS with zkSNARKs (Zero-Knowledge Succinct Non-Interactive Arguments of Knowledge), we provide a framework for privacy-preserving and verifiable operations in Cyber-Physical Systems (CPS). We demonstrate how SQL (Structured Query Language) queries can be transformed into relational algebra and compiled into ZoKrates Domain-Specific Language (DSL) for ZKP generation. A case study in sustainable food supply chains highlights the approach's potential to enhance trust, scalability, and security in collaborative environments.
Alex Wong, Duncan McFarlane, Charlotte Ellarby, M.B. Lee · 5 authors
Twenty-five years ago, the specification of the Intelligent Product was established, envisaging real-time connectivity that not only enables products to gather accurate data about themselves but also allows them to assess and influence their own destiny. Early work by the Auto-ID project focused on creating a single, open-standard repository for storing and retrieving product information, laying a foundation for scalable connectivity. A decade later, the approach was revisited in light of low-cost RFID systems that promised a low-cost link between physical goods and networked information environments. Since then, advances in blockchain, Web3, and artificial intelligence have introduced unprecedented levels of resilience, consensus, and autonomy. By leveraging decentralised identity, blockchain-based product information and history, and intelligent AI-to-AI collaboration, this paper examines these developments and outlines a new specification for the Intelligent Product 3.0, illustrating how decentralised and AI-driven capabilities facilitate seamless interaction between physical AI and everyday products.
The emergence of AI agents and agentic systems represents a significant milestone in artificial intelligence, enabling autonomous systems to operate, learn, and collaborate in complex environments with minimal human intervention. This paper, drawing on multi-expert perspectives, examines the potential of AI agents and agentic systems to reshape industries by decentralizing decision-making, redefining organizational structures, and enhancing cross-functional collaboration. Specific applications include healthcare systems capable of creating adaptive treatment plans, supply chain agents that predict and address disruptions in real-time, and business process automation that reallocates tasks from humans to AI, improving efficiency and innovation. However, the integration of these systems raises critical challenges, including issues of attribution and shared accountability in decision-making, compatibility with legacy systems, and addressing biases in AI-driven processes. The paper concludes that while agentic systems hold immense promise, robust governance frameworks, cross-industry collaboration, and interdisciplinary research into ethical design are essential. Future research should explore adaptive workforce reskilling strategies, transparent accountability mechanisms, and energy-efficient deployment models to ensure ethical and scalable implementation.
Abstract This paper introduces semi-competitive differential game logic $$\textsf {dG}\mathcal {L}_{sc}$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:mrow> <mml:mi>dG</mml:mi> <mml:msub> <mml:mi>L</mml:mi> <mml:mrow> <mml:mi>sc</mml:mi> </mml:mrow> </mml:msub> </mml:mrow> </mml:math> , which enables verification of safety-critical applications that involve interactions between two agents. In $$\textsf {dG}\mathcal {L}_{sc}$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:mrow> <mml:mi>dG</mml:mi> <mml:msub> <mml:mi>L</mml:mi> <mml:mrow> <mml:mi>sc</mml:mi> </mml:mrow> </mml:msub> </mml:mrow> </mml:math> , these interactions are specified as games on hybrid systems with two players that may collaborate with each other when helpful and may compete when necessary. The players in the hybrid games of $$\textsf {dG}\mathcal {L}_{sc}$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:mrow> <mml:mi>dG</mml:mi> <mml:msub> <mml:mi>L</mml:mi> <mml:mrow> <mml:mi>sc</mml:mi> </mml:mrow> </mml:msub> </mml:mrow> </mml:math> have individual goals that may overlap, leading to nonzero-sum games. This makes $$\textsf {dG}\mathcal {L}_{sc}$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:mrow> <mml:mi>dG</mml:mi> <mml:msub> <mml:mi>L</mml:mi> <mml:mrow> <mml:mi>sc</mml:mi> </mml:mrow> </mml:msub> </mml:mrow> </mml:math> especially well-suited for verifying situations where players, e.g., share safety objectives but otherwise pursue different goals, so that zero-sum assumptions lead to overly conservative results. Additionally, $$\textsf {dG}\mathcal {L}_{sc}$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:mrow> <mml:mi>dG</mml:mi> <mml:msub> <mml:mi>L</mml:mi> <mml:mrow> <mml:mi>sc</mml:mi> </mml:mrow> </mml:msub> </mml:mrow> </mml:math> solves the subtlety that even though each player may benefit from knowledge of the other player’s goals, e.g., concerning shared safety objectives, unsafe situations might still occur if every player were to mutually assume the other player would act to avoid unsafety. The syntax and semantics, as well as a sound and relatively complete proof calculus are presented for $$\textsf {dG}\mathcal {L}_{sc}$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:mrow> <mml:mi>dG</mml:mi> <mml:msub> <mml:mi>L</mml:mi> <mml:mrow> <mml:mi>sc</mml:mi> </mml:mrow> </mml:msub> </mml:mrow> </mml:math> . The relationship between $$\textsf {dG}\mathcal {L}_{sc}$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:mrow> <mml:mi>dG</mml:mi> <mml:msub> <mml:mi>L</mml:mi> <mml:mrow> <mml:mi>sc</mml:mi> </mml:mrow> </mml:msub> </mml:mrow> </mml:math> and zero-sum differential game logic $$\textsf {dG}\mathcal {L}$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:mrow> <mml:mi>dG</mml:mi> <mml:mi>L</mml:mi> </mml:mrow> </mml:math> is discussed and the purpose of $$\textsf {dG}\mathcal {L}_{sc}$$ <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML"> <mml:mrow> <mml:mi>dG</mml:mi> <mml:msub> <mml:mi>L</mml:mi> <mml:mrow> <mml:mi>sc</mml:mi> </mml:mrow> </mml:msub> </mml:mrow> </mml:math> illustrated in a canonical example.
Lukas Kirchdorfer, Robert Blümel, Timotheus Kampik, Han van der Aa · 5 authors
Abstract Business process simulation (BPS) is a powerful tool for estimating process performance across different scenarios, offering critical support for organizational process redesign and optimization. Traditional BPS approaches predominantly rely on a control-flow-first perspective by enriching a process model with simulation parameters. While these approaches seem suitable for capturing centrally orchestrated processes, such as those managed by workflow systems, they fall short of accurately reflecting real-world processes characterized by decentralized decision-making and distinct resource behaviors. To overcome this limitation, we propose AgentSimulator , a resource-first BPS approach that discovers a multi-agent system from an event log. By modeling the distinct behaviors and interaction patterns of individual resources, AgentSimulator effectively simulates the underlying process. Our approach automatically identifies whether resource behavior is rather orchestrated or autonomous, adapting to the specific decision-making structure of the process. Experimental results reveal that AgentSimulator achieves state-of-the-art simulation accuracy while ensuring high adaptability to various process types.
Sergii Grybniak, Yevhen Leonchyk, Igor Y. Mazurok, Oleksandr S. Nashyvan · 6 authors
This paper presents an approach to improving the scalability of the decentralized smart contract platform Waterfall, based on the concept of hierarchical fractal sharding. Although distributed ledger technology holds significant promise for building secure and transparent digital ecosystems, its widespread adoption remains limited by scalability issues. A key challenge lies in the inability to proportionally increase transaction throughput with the growing number of participants without undermining either decentralization or security. The proposed solution reduces both computational and communication loads by distributing transactions, smart contracts, and network state across a system of recursively structured shards. Each shard operates as an independently validated subnetwork organized as a directed acyclic graph structure that supports asynchronous execution and consensus. This design enables the participation of low-power nodes, enhances load balancing, and achieves scalability not only at the level of the entire network but also within its internal components. The study details the mechanisms for shard formation and merging, transaction routing strategies, and dynamic placement of smart contracts. In addition, a probabilistic model is introduced to evaluate the risk of malicious capture of individual shards, and guidelines are provided for choosing safe shard sizes under various threat assumptions. While the proposed architecture is designed specifically for the Waterfall platform, its core principles and several of its methods may be adapted to other distributed ledger systems, including but not limited to blockchain-based platforms, particularly those employing modular or directed acyclic graph-structured architectures.
Blockchain consensus mechanisms have relied on algorithms such as Proof-of-Work (PoW) and Proof-of-Stake (PoS) to ensure network functionality and integrity. However, these approaches struggle with adaptability for decision-making where the opinions of each matter rather than reaching an agreement based on honest majority or weighted consensus. This paper introduces a novel deliberation-based consensus mechanism where Large Language Models (LLMs) act as rational agents engaging in structured discussions to reach a unanimous consensus. By leveraging graded consensus and a multi-round deliberation process, our approach ensures unanimous consensus for definitive problems and graded consensus for prioritized decision problems and policies. We provide a formalization of our system and use it to show that the properties of blockchains are maintained, while also addressing the behavior in terms of adversaries, stalled deliberations, and confidence in consensus. Moreover, experimental results demonstrate system feasibility, showcasing convergence, block properties, and accuracy, which enable deliberative decision-making on blockchain networks.
The rapid advancement of Large Language Models (LLMs) has catalyzed the development of multi-agent systems, where multiple LLM-based agents collaborate to solve complex tasks. However, existing systems predominantly rely on centralized coordination, which introduces scalability bottlenecks, limits adaptability, and creates single points of failure. Additionally, concerns over privacy and proprietary knowledge sharing hinder cross-organizational collaboration, leading to siloed expertise. To address these challenges, we propose AgentNet, a decentralized, Retrieval-Augmented Generation (RAG)-based framework that enables LLM-based agents to autonomously evolve their capabilities and collaborate efficiently in a Directed Acyclic Graph (DAG)-structured network. Unlike traditional multi-agent systems that depend on static role assignments or centralized control, AgentNet allows agents to specialize dynamically, adjust their connectivity, and route tasks without relying on predefined workflows. AgentNet’s core design is built upon several key innovations: (1) Fully Decentralized Paradigm: Removing the central orchestrator, allowing agents to coordinate and specialize autonomously, fostering fault tolerance and emergent collective intelligence. (2) Dynamically Evolving Graph Topology: Real-time adaptation of agent connections based on task demands, ensuring scalability and resilience. (3) Adaptive Learning for Expertise Refinement: A retrieval-based memory system that enables agents to continuously update and refine their specialized skills. By eliminating centralized control, AgentNet enhances fault tolerance, promotes scalable specialization, and enables privacy-preserving collaboration across organizations. Through decentralized coordination and minimal data exchange, agents can leverage diverse knowledge sources while safeguarding sensitive information. Experimental results demonstrate that AgentNet outperforms traditional centralized multi-agent systems, significantly improving efficiency, adaptability, and scalability in dynamic environments, making it a promising foundation for next-generation autonomous, privacy-respecting multi-agent ecosystems.
This article examines the transformative impact of Agentic Process Automation (APA) on modern business workflows, highlighting the evolution from traditional Robotic Process Automation to autonomous intelligent systems. The article establishes APA as a paradigm shift that transcends the limitations of conventional automation approaches through self-governing agent models capable of adaptive decision-making. Through comprehensive analysis spanning architectural foundations, comparative capabilities, multi-agent collaboration frameworks, and real-world implementations, this article demonstrates how APA systems deliver superior performance in dynamic business environments. Key aspects explored include decentralized intelligence, machine learning integration, ethical governance frameworks, and strategic implementation methodologies. Case studies across financial services, healthcare, and manufacturing sectors provide empirical evidence of APA's operational benefits, while also highlighting implementation challenges and mitigation strategies. The article reveals that organizations implementing agentic systems achieve significant improvements in process efficiency, adaptability, and cost optimization compared to traditional automation approaches, particularly for complex workflows requiring judgment and contextual understanding. This article provides valuable insights for organizations navigating the transition toward intelligent automation and offers a structured framework for evaluating APA readiness, implementation priorities, and governance considerations within enterprise environments
Luigi Bellomarini, Marco Favorito, Giuseppe Galano
This work proposes a novel formal verification technique to analyze Bitcoin smart contracts (when specified in BITML) through ATL model checking, using the MCMAS model checker. In particular, we developed a translation procedure from a BITML contract to a MCMAS model that simulates the BITML semantics, hence allowing for strategic reasoning on BITML smart contracts. We implemented the technique in a prototype tool, which we tested over several case studies, showing that we can verify smart contract specifications that capture interesting multi-agent interactions and strategic specifications.
While the trend of decentralized governance is obvious (cryptocurrencies and blockchains are widely adopted by multiple sovereign countries), initiating governance proposals within Decentralized Autonomous Organizations (DAOs) is still challenging, i.e., it requires providing a low-level transaction payload, therefore posing significant barriers to broad community participation. To address these challenges, we propose a multi-agent system powered by Large Language Models with a novel Label-Centric Retrieval algorithm to automate the translation from natural language inputs into executable proposal transactions. The system incorporates DAOLang, a Domain-Specific Language to simplify the specification of various governance proposals. The key optimization achieved by DAOLang is a semantic-aware abstraction of user input that reliably secures proposal generation with a low level of token demand. A preliminary evaluation on real-world applications reflects the potential of DAOLang in terms of generating complicated types of proposals with existing foundation models, e.g. GPT-4o.
Decentralized Autonomous Organizations (DAOs) are a type of Decentralized Applications (DApps) that utilize smart contracts to support governance processes. To achieve a high degree of utility of the system, stakeholders need to identify a suitable organizational structure in the early stages of design. While Model-Driven Development (MDD) methods are established for DApp and smart contract design, they lack specialization for modeling the organizational structures of DAOs. To address this gap, we propose a modeling language and a method which support crucial DAO design and development phases. The method is evaluated through an in vivo case study. Unlike existing solutions, comprehensive stepwise guidance is provided by our method for both technical and non-technical stakeholders involved in DAO development from the initial stages of the project.
As blockchain technology continues to evolve, the need for accessible solutions for developing smart contracts has grown, especially for non-technical users. This paper addresses practitioners' challenges in generating Solidity smart contracts from natural language requirements within the AstraKode Blockchain no-code platform (AKB). Our goal is to lower the barrier of entry into smart contract development, making it more accessible to users with limited technical expertise. We propose three methods, i.e., Naive Generation, Augmented Generation, and Enhanced Generation, each utilizing large language models to streamline the code generation process. These methods cater to different user needs, from rapid prototyping to handling complex business scenarios, improving accessibility and usability within AKB. We demonstrate their practical relevance, potential, and limitations in addressing real-world challenges in smart contract development through empirical evaluations and practitioner feedback. Thanks to collaboration with academia and effective knowledge transfer, these methods provide innovative solutions to the challenges of smart contract generation. Furthermore, they have been integrated into AKB to enhance user services, ultimately promoting the development and deployment of secure and efficient smart contracts in the industry.
Artificial Intelligence in Law
Advanced Research in Systems and Signal Processing
Jun Huang, Yonglin Tian, Tengchao Zhang, Qinghua Ni · 8 authors
ABSTRACT Tourism is a critical driver of economic growth and cultural exchange. However, traditional tourism models often struggle to address challenges such as insufficient understanding of diverse travel preferences, limited capacity for personalized services, and a lack of intelligent, convenient service solutions. In this paper, we propose a virtual‐real interactive Parallel Tourism System (PTS) and a Smart Tourism and Lifestyle Service Large Model (STLS‐LM) to handle these challenges. By integrating Decentralized Autonomous Organizations (DAOs) and Foundation Models (FMs) supported by Retrieval‐Augmented Generation, PTS elevates tourism services by incorporating agentic intelligence and autonomous intelligence, advancing beyond traditional AI‐powered applications. The proposed framework enables a deep understanding of task requirements, precise multi‐source information sensing, personalized functionality customization, and timely public feedback adoption, forming a comprehensive intelligent travel service system that caters to the evolving needs of modern tourism.
Decompiler is a specialized type of reverse engineering tool extensively employed in program analysis tasks, particularly in program comprehension and vulnerability detection. However, current Solidity smart contract decompilers face significant limitations in reconstructing the original source code. In particular, the bottleneck of SOTA decompilers lies in inaccurate function identification, incorrect variable type recovery, and missing contract attributes. These deficiencies hinder downstream tasks and understanding of the program logic. To address these challenges, we propose SmartHalo, a new framework that enhances decompiler output by combining static analysis (SA) and large language models (LLM). SmartHalo leverages the complementary strengths of SA’s accuracy in control and data flow analysis and LLM’s capability in semantic prediction. More specifically, SmartHalo constructs a new data structure - Dependency Graph (DG), to extract semantic dependencies via static analysis. Then, it takes DG to create prompts for LLM optimization. Finally, the correctness of LLM outputs is validated through symbolic execution and formal verification. Evaluation on a dataset consisting of 465 randomly selected smart contract functions shows that SmartHalo significantly improves the quality of the decompiled code, compared to SOTA decompilers (e.g., Gigahorse). Notably, integrating GPT-4o mini with SmartHalo further enhances its performance, achieving a precision of 91.32% and a recall of 87.38% for function boundaries, a precision of 90.40% and a recall of 88.82% for variable types, and a precision of 80.66% and a recall of 91.78% for contract attributes.
This paper presents a comprehensive Model-Driven Engineering (MDE) methodology for automatically transforming Business Process Model and Notation (BPMN) diagrams into executable blockchain-based smart contracts. The proposed approach defines a set of Atlas Transformation Language (ATL) rules that systematically map BPMN elements to Solidity con-structs, ensuring semantic consistency and traceability through-out the transformation process. The framework integrates several stages, including process modeling, model validation, code generation, and deployment, supported by tools such as Camunda, Eclipse ATL, Remix IDE, and MetaMask. Experimental vali-dation on the Ethereum Sepolia test network demonstrates the approach’s ability to enhance automation, reduce manual coding errors, and improve synchronization between business work-flows and their on-chain implementations. Compared to existing BPMN-to-blockchain frameworks, the proposed solution offers a unified and reusable transformation pipeline that bridges the gap between business process modeling and blockchain execution. The study concludes that MDE provides a scalable, traceable, and standardized foundation for developing decentralized business process applications.
Elizaveta Pertseva, Alex Ozdemir, Shankara Pailoor, Alp Bassa · 7 authors
Abstract This paper presents a new refutation procedure for multimodular systems of integer constraints that commonly arise when verifying cryptographic protocols. These systems, involving polynomial equalities and disequalities modulo different constants, are challenging for existing solvers due to their inability to exploit multimodular structure. To address this issue, our method partitions constraints by modulus and uses lifting and lowering techniques to share information across subsystems, supported by algebraic tools like weighted Gr bner bases. Our experiments show that the proposed method outperforms existing state-of-the-art solvers in verifying cryptographic implementations related to Montgomery arithmetic and zero-knowledge proofs.
Abstract Decentralized autonomous organizations (DAOs), like blockchains, were expected to revolutionize the economy by solving the principal–agent problem in the corporate form. Yet, the very first DAO (“The DAO”) suffered a catastrophic hack that led to the Ethereum blockchain’s forking. Despite this failure, and the 2022 crypto markets collapse, enthusiasm for DAOs as corporate alternatives persists in crypto circles. Nine years on, however, DAOs remain a niche structure with little traction outside crypto projects. This chapter posits that this is due to fundamental flaws in the case for DAOs as corporate replacements. Crypto proponents often dismiss centralization and opacity as wholly undesirable, overlooking their functional value in business organizations. They also fail to recognize that blockchain’s security, transparency, and predictability are narrowly bound and come at a cost, rendering DAOs poorly equipped to handle the risks organizations face in the real world. These trade-offs explain why DAOs are less a revolution, more naïve wishful thinking.
S. Gheisari, Jaime Osvaldo Salas, Semih Yumuşak, George Konstantinidis
This paper presents MINiDM, a novel negotiation framework designed for decentralised data marketplaces. MINiDM enables AI-driven agents to negotiate complex, multi-issue data-sharing agreements while ensuring compliance with legal and ethical standards such as GDPR. The framework combines game-theoretic strategies, formal policy languages and vocabularies (ODRL and DPV), and a lightweight multi-agent commitment protocol to automate fair transactions without relying on central authorities or blockchain- based contracts. Experimental results demonstrate that MINiDM outperforms existing protocols in terms of agreement rate, fairness, and negotiation efficiency, offering a promising direction for secure and autonomous data trading in Web3 ecosystems.