引用本文: |
-
李广,陈梓钿,卞静,周杰英,吴维刚.区块链欺诈行为识别技术综述[J].信息安全学报,2024,9(4):1-30 [点击复制]
- LI Guang,CHEN Zitian,BIAN Jing,ZHOU Jieying,WU Weigang.Blockchain Fraud Behaviors Detection Technology: A Survey[J].Journal of Cyber Security,2024,9(4):1-30 [点击复制]
|
|
|
|
本文已被:浏览 4143次 下载 2615次 |
![本文二维码信息](/uploadfile/two_dimensional_code_img/xxaqxb/2024/4/20240401.png) 码上扫一扫! |
区块链欺诈行为识别技术综述 |
李广, 陈梓钿, 卞静, 周杰英, 吴维刚
|
|
(中山大学 计算机学院 广州 中国 510006) |
|
摘要: |
近年来, 区块链技术和产业的迅速发展为经济和技术创新注入了新的活力, 但随之而来的是不断涌现的欺诈行为。这些欺诈行为不仅对用户造成了经济损失, 也对区块链技术的信誉和发展带来了威胁。因此, 识别和预防欺诈行为对于保障区块链技术和产业创新的良性发展至关重要。另一方面, 区块链欺诈行为变化快、匿名性强, 具有多样性和复杂性, 给识别工作带来了巨大挑战, 针对这些挑战, 目前已提出了相当多的技术方法。 本文整理归纳了近五年来的相关文献, 清晰呈现区块链欺诈行为识别技术的最新进展。考虑到识别技术的多样性, 本文采用了两层的分类框架对其进行归纳。首先从业务场景出发, 划分出不同类型的欺诈行为, 涉及区块链洗钱、非法代币发行、庞氏骗局和钓鱼诈骗等八种行为。进而, 再针对每一类欺诈行为, 分析讨论对应的识别技术。通过对识别技术解析与归纳, 本文将识别技术从具体场景中抽象出来, 构建出一般化的识别技术体系。并基于这一体系对识别技术展开详细讨论, 包括: 区块链交易图构建技术、特征工程方法以及欺诈行为识别方法与模型。在识别方法上, 本文重点关注了近年流行的区块链去中心化生态下的一些反欺诈识别工作, 包括: 跨链洗钱识别、 去中心化平台的代币骗局识别等, 此类欺诈行为具有较高的复杂性和识别难度, 与之相关的识别技术还较少, 亟待得到更多的关注。最后, 本文依据当前区块链欺诈行为识别工作所面临的挑战和困难, 分析了未来技术趋势。 |
关键词: 区块链 反欺诈 欺诈识别 钓鱼识别 洗钱识别 |
DOI:10.19363/J.cnki.cn10-1380/tn.2024.07.01 |
投稿时间:2022-10-17修订日期:2023-02-28 |
基金项目:本课题得到广东省重点领域研发计划项目(No. 2020B0101090005)资助。 |
|
Blockchain Fraud Behaviors Detection Technology: A Survey |
LI Guang, CHEN Zitian, BIAN Jing, ZHOU Jieying, WU Weigang
|
(School of Computer Science and Engineering, Sun Yat-Sen University, Guangzhou 510006, China) |
Abstract: |
In recent years, the rapid development of blockchain technology and industry has engendered fresh vigor for economic and technological innovation. However, this also catalyzed the proliferation of fraudulent practices, which have posed a threat to the technology's reputation and progression as well as brings financial loss to subscribers. As a result, it is crucial to detect and prevent fraud behaviors to safeguard the benign development of blockchain technology and application. On the other hand, blockchain fraud behaviors is anonymous, diverse and complex, which poses enormous challenges to the detection methods, and quite many different approaches have been proposed to address these challenges. This paper summarizes the relevant literature in the past five years, to clearly present the latest progress of blockchain fraud behavior detection technologies. In consideration of the diversity of detection technologies, this paper adopts a two-level classification framework. Firstly, the fraud behaviors are classified into different categories based on business scenarios, involving blockchain money laundering, initial coin offering, Ponzi schemes and phishing scams, and totally eight types behaviors. And then, for each category, the corresponding detection techniques are analyzed and discussed. By analyzing and summarizing, this paper views the detection technologies from specific scenarios and constructs a generalized technical architecture. The detection technologies are discussed in detail based on this architecture, covering blockchain transaction graph construction technologies, feature engineering technologies, fraud behavior detection technologies and models. This paper also discusses the fraud behavior detection efforts in the decentralized ecosystem of blockchain, including cross-chain money laundering, decentralized platform token scam, etc, which has become popular in recent years. These fraud behaviors have high complexity and detection difficulty, and related detection technologies are still lacking, which urgently require more attention. Finally, this paper analyzes the future technology trends based on the challenges and difficulties faced by the current blockchain fraud behavior detection work. |
Key words: blockchain Anti-Fraud fraud detection money laundering detection phishing detection |
|
|
|
|
|