The Fourth International Workshop on Smart Data for Blockchain and Distributed Ledger (SDBD’24)

Co-located with ACM SIGKDD 2024, August 26th, Barcelona, Spain

Oral Presentation Papers

Enabling Encrypted Consensus through Fully Holomorphic Encryption in Proof of Intelligence Mechanisms

Dennis Song, Yuping Yan, George Shao and Mason Song
Abstract: Ensuring privacy and security while reaching consensus presents major issues in evolving decentralized systems. Traditional consensusbuilding techniques without encryption frequently compromise privacy and confidentiality by requiring participants to reveal sensitive information. Meanwhile, with the integration of Artificial Intelligence (AI) into Web3, these traditional consensuses, such as Proof of Work (PoW) and Proof of Stake (PoS) can not fit the characteristics of AI tasks anymore. To address these issues, we propose a novel approach leveraging Fully Homomorphic Encryption (FHE) to encrypt Proof of Intelligence (PoI) consensus in AI tasks. This framework guarantees data privacy with the calculations on encrypted data during the consensus procedure. Additionally, it mitigates the risk of malicious clients stealing or free-riding on others’ value within the decentralized system. By proving the security properties of correctness and ciphertext indistinguishability with FHE, our approach offers a robust solution for achieving consensus while safeguarding the integrity and confidentiality of participants’ data. To validate our framework, we demonstrate its effectiveness through empirical security analysis and performance analysis.

Benchmarking GNNs Using Lightning Network Data

Rainer Feichtinger, Florian Grötschla, Lioba Heimbach and Roger Wattenhofer
Abstract: The Bitcoin Lightning Network is a layer 2 protocol designed to facilitate fast and inexpensive Bitcoin transactions. It operates by establishing channels between users, where Bitcoin is locked and transactions are conducted off-chain until the channels are closed, with only the initial and final transactions recorded on the blockchain. Routing transactions through intermediary nodes is crucial for users without direct channels, allowing these routing nodes to collect fees for their services. Nodes announce their channels to the network, forming a graph with channels as edges. In this paper, we analyze the graph structure of the Lightning Network and investigate the statistical relationships between node properties using machine learning, particularly Graph Neural Networks (GNNs). We formulate a series of tasks to explore these relationships and provide benchmarks for GNN architectures, demonstrating how topological and neighbor information enhances performance. Our evaluation of several models reveals the effectiveness of GNNs in these tasks and highlights the insights gained from their application.

Poster Papers

Web3Env: An Open-source Interactive Platform for Scientific Research and Applications in Web3

Zesen Zhuang, Xinyu Tian and Luyao Zhang
Abstract: The emergence of Web3 has provided new opportunities but still lacks technology support for research and innovation. We propose Web3Env, the first open-source online platform for Web3 scientific research and applications. The design of Web3Env is intended for interdisciplinary collaborations and human-centered computation. Firstly, Web3 provides a reinforcement learning environment that can be customized for scientific research in various application scenarios in Web3. Secondly, Web3Env offers open APIs that support interaction with reinforcement learning data and algorithms, connecting the Web3-oriented environments to specific computational tasks. Last but not least, Web3Env supports non-coding programming by providing a user interface (UI), facilitating collaborations of professionals and experts from a broad range of disciplines. We present the design, demo, and two applications of our platform for Web3.

Unlocking the Economics of Decentralized Physical Infrastructure Networks: Taxonomy, Data, and Impact Data

Luyao Zhang and Yulin Liu
Abstract: Decentralized Physical Infrastructure Network (DePIN) represents an innovative approach to managing physical infrastructure using blockchain technology. By 2024, the market capitalization of DePIN-related projects has surpassed $10 billion, reflecting significant popularity and investment. Despite the excitement within the Web3 community, there is a notable lack of academic research on DePIN’s foundational aspects and economic impact. This paper addresses these gaps with a comprehensive analysis of the DePIN ecosystem. We categorize various DePIN networks by economic sectors, examine the unique characteristics of DePIN data assets, and investigate the economic mechanisms of DePIN implementations. Through a detailed taxonomy, real-world data analysis, and assessment of economic incentives, this study aims to enhance the understanding of DePIN and to support its development as a sustainable and efficient model for managing physical infrastructure.

Analyzing the Blockchain Trilemma: Performance Metrics in Algorand and Ethereum 2.0

Mingwei Jing, Yihang Fu, Jiaolun Zhou, Ye Wang, Luyao Zhang and Chuang Hu
Abstract: Blockchain technology is critical to the future of the digital economy and the metaverse, serving as a foundation for everything from decentralized finance to virtual assets. Yet, its full potential is constrained by the "Blockchain Trilemma," which necessitates a balance among decentralization, security, and scalability. This study compares Algorand and Ethereum 2.0, two leading blockchain platforms, to evaluate how they measure up against these critical metrics. Our research analyzes decentralization indices, scalability metrics, and security protocols within both ecosystems. The findings aim to shed light on each platform’s strategies and their efficacy in overcoming the trilemma challenges, thus advancing the dialogue on blockchain development for robust digital infrastructures.

Model Agnostic Hybrid Sharding for Heterogeneous Distributed Inference

Claudio Angione, Yue Zhao, Harry Yang, Ahmad Farhan, Fielding Johnston, James Buban and Patrick Colangelo
Abstract: The rapid growth of large-scale AI models, particularly large language models has brought significant challenges in data privacy, computational resources, and accessibility. Traditional centralized architectures often struggle to meet required data security and scalability needs which hinders the democratization of AI systems. Nesa introduces a model-agnostic sharding framework designed for decentralized AI inference. Our framework uses blockchain-based sequential deep neural network sharding to distribute computational tasks across a diverse network of nodes based on a personalised heuristic and routing mechanism. This enables efficient distributed training and inference for recent large-scale models even on consumer-grade hardware. We use compression techniques like dynamic blockwise quantization and mixed matrix decomposition to reduce data transfer and memory needs. We also integrate robust security measures, including hardware-based trusted execution environments to ensure data integrity and confidentiality. Evaluating our system across various natural language processing and vision tasks shows that these compression strategies do not compromise model accuracy. Our results highlight the potential to democratize access to cutting-edge AI technologies by enabling secure and efficient inference on a decentralized network.

Complete Security and Privacy for AI Inference in Decentralized Systems

Hongyang Zhang, Yue Zhao, Claudio Angione, Harry Yang, James Buban, Ahmad Farhan, Fielding Johnston and Patrick Colangelo
Abstract: The need for data security and model integrity has been accentuated by the rapid adoption of AI and ML in data-driven domains including healthcare, finance, and security. Large models are crucial for tasks like diagnosing diseases and forecasting finances but tend to be delicate and not very scalable. Decentralized systems solve this issue by distributing the workload and reducing central points of failure. Yet, data and processes spread across different nodes can be at risk of unauthorized access, especially when they involve sensitive information. Nesa solves these challenges with a comprehensive framework using multiple techniques to protect data and model outputs. This includes zero-knowledge proofs for secure model verification. The framework also introduces consensus-based verification checks for consistent outputs across nodes and confirms model integrity. Split Learning divides models into segments processed by different nodes for data privacy by preventing full data access at any single point. For hardware-based security, trusted execution environments are used to protect data and computations within secure zones. Nesa’s state-of-the-art proofs and principles demonstrate the framework’s effectiveness, making it a promising approach for securely democratizing artificial intelligence.