ART: A Graph-based Framework for Investigating Illicit Activity in Monero via Address-Ring-Transaction Structures

Andrea Venturi, Imanol Jerico-Yoldi, Francesco Zola, Raul Orduna · 2025

Read ART: A Graph-based Framework for Investigating Illicit Activity in Monero via Address-Ring-Transaction Structures by Andrea Venturi, Imanol Jerico-Yoldi, Francesco Zola, Raul Orduna (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-16192.txt.

Introduction

Andrea Venturi, Imanol Jerico-Yoldi, Francesco Zola, and Raul Orduna build a graph frame to probe illicit activity on the Monero network. Monero is a privacy coin whose strong untraceability defeats standard blockchain analysis. Law enforcement still needs to grasp criminal behavior within it. Their case study starts from transactions already linked to found criminal activity.

Address, ring, and transaction data are built into graphs that capture Monero ring signature structure. Structural and temporal features from those graphs train machine learning models. The models detect similar behavior patterns that may reveal criminal methods. The work is a first partial step toward tools for investigators in privacy-preserving blockchain systems.

Catalog record

CREATORAndrea Venturi, Imanol Jerico-Yoldi, Francesco Zola, Raul Orduna
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE21 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2511-16192

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

ART: A Graph-based Framework for Investigating Illicit Activity in Monero via Address-Ring-Transaction Structures Andrea Venturi, Imanol Jerico-Yoldi, Francesco Zola, Raul Orduna Abstract As Law Enforcement Agencies advance in cryptocurrency forensics, criminal actors aiming to conceal illicit fund movements increasingly turn to “mixin” services or privacy-based cryptocurrencies. Monero stands out as a leading choice due to its strong privacy preserving and untraceability properties, making conventional blockchain analysis ineffective. Understanding the behavior and operational patterns of criminal actors within Monero is therefore challenging and it is essential to support future investigative strategies and disrupt illicit activities. In this work, we propose a case study in which we leverage a novel graph-based methodology to extract structural and temporal patterns from Monero transactions linked to already discovered criminal activities. By building Address-Ring-Transaction graphs from flagged transactions, we extract structural and temporal features and use them to train Machine Learning models capable of detecting similar behavioral patterns that could highlight criminal modus operandi. This represents a first partial step toward developing analytical tools that support investigative efforts in privacy-preserving blockchain ecosystems. I Introduction The decentralized and pseudo-anonymous nature of cryptocurrencies has made them a favored tool for organized criminal groups seeking to conceal the flow of illicit funds. In response, Law Enforcement Agencies (LEAs) have increasingly invested in blockchain forensics, developing techniques to analyze and disrupt the anonymity level provided by transparent blockchains like Bitcoin. These advances have driven criminal actors to adopt further obfuscation strategies, including the use of crypto-mixers [3] and a growing shift toward privacy-focused cryptocurrencies. Among these, Monero has emerged as one of the most widely used and technically advanced privacy coins. Launched in April 2014, Monero is specifically designed to provide strong privacy guarantees through its RingCT protocol, which ensures both untraceability and unlinkability of transactions [ 6 ] . Its adoption has been documented in several criminal contexts. For instance, the January 2024 Chainalysis report highlights the recent shift of Child Sexual Abuse Material (CSAM) vendors towards instant exchanges offering Monero [ 2 ] . Similarly, Monero has been linked to high-profile cybercrime operations, including the Lazarus Group’s to launder proceeds from the WannaCry 2.0 ransomware campaign [ 1 ] , and LockBit Group for gathering the ransom after their attacks [ 9 ] . This work takes an initial step toward understanding and characterizing the modus operandi of criminals on the Monero blockchain. While previous research has primarily focused on developing traceability methods or identifying security vulnerabilities to compromise Monero’s cryptographic protections (e.g., [ 5 , 4 ] ), our approach shifts the focus toward behavioral analysis. We introduce a graph-based method that models Monero’s transaction ecosystem through the Address-Ring-Transaction (ART) graph, a novel representation specifically designed to capture unique properties of Monero’s privacy-preserving protocol. This representation enables the extraction of structural and temporal features from transactions associated with specific behaviors, which can then be used to train Machine Learning (ML) classifiers capable of identifying other transactions exhibiting similar patterns within the blockchain. We validate our methodology through a case study based on real Monero transactions linked to illicit activities, demonstrating the potential of behavioral graph-based analysis as a complementary means for investigating privacy-oriented cryptocurrencies. By focusing on behavioral patterns rather than protocol-level vulnerabilities, this approach uncovers transactional regularities that can guide future detection strategies and pave the way for advanced investigative tools in privacy-preserving blockchain systems II Background and Problem Statement Monero’s privacy model is primarily based on the Ring Confidential Transactions (RingCT) protocol [ 6 ] , which combines three fundamental privacy-enhancing technologies: ring signatures, stealth addresses, and confidential transactions. Ring signatures enable a sender to obscure their identity within a group of potential signers. In Monero, every transaction input is mixed with decoy inputs (also called mixins) from the blockchain, making it computationally infeasible to determine which input is the actual spender. Stealth addresses enable recipients to receive funds via one-time addresses generated from their public key, ensuring that transactions cannot be linked back to their actual address. Confidential transactions use Pedersen commitments to conceal transaction amounts. This prevents third parties from viewing or analyzing transferred values while still allowing network participants to verify transaction validity. Together, these technologies make Monero a preferred tool by individuals who aim to obscure their financial movements, including actors involved in illicit activities, such as money laundering, ransomware payments and darknet market transactions. As a consequence, understanding and modeling the behavior and modus operandi of criminals leveraging Monero is thus a critical task. Our goal in this work is to provide a first analytical method that could support LEAs in detecting common patterns exhibited by criminals when performing obfuscation operations in Monero. III Related Work A significant portion of the existing literature on Monero focuses on transaction traceability by developing heuristic methods aimed at identifying the true input among ring signatures, rather than examining potential criminal modus operandi. Seminal works in this area, such as [ 5 ] and [ 10 ] , introduced heuristic-based attacks capable of significantly narrowing the set of potential senders, thereby undermining anonymity. However, these techniques were only effective against early versions of the Monero protocol and are no longer applicable. More recent approaches have been proposed [ 2 ] , but their effectiveness is limited, and they typically require stringent conditions and considerable computational resources. In contrast, our work addresses the problem from a different angle, focusing on behavioral analysis within the Monero blockchain. This offers a novel and complementary perspective to the current state of the art. While our graph-based methodology is inspired by prior research on behavioral analysis (e.g., [ 1 ] ), to the best of our knowledge, this is the first time such an approach has been applied to Monero, and previous studies have generally focused on traditional, transparent cryptocurrencies (e.g., [ 3 , 7 , 8 ] ). IV Behavior Detection Method We introduce a novel graph-based method to uncover behavioral patterns and identify specific—potentially illicit—modus operandi within the Monero blockchain. Unlike transparent cryptocurrencies such as Bitcoin, Monero’s strong privacy guarantees make direct observation of user activity infeasible, posing a fundamental challenge to traditional blockchain analysis. Our approach addresses this limitation by modeling the Monero network through a new graph representation that captures both its structural and temporal dynamics. Our approach is divided into four main phases which are detailed below. IV-A Data Acquisition The main objective of the proposed methodology is to train binary classifiers capable of detecting transaction patterns that resemble a given modus operandi inside the Monero blockchain. A crucial step in this process is the acquisition of sufficient labeled data that accurately represents the behavioral characteristics of the target activity. Due to the privacy mechanisms implemented in Monero—such as one-time stealth addresses and ring signatures—traditional address-based tracing is not feasible. Therefore, our analysis focuses on transactions themselves, which constitute the most accessible and reliable source of information within Monero’s privacy framework. By obtaining a set of transactions labeled according to the modus operandi of interest, we can observe and reconstruct the associated transaction flows, enabling the identification of distinct behavioral patterns linked to specific entities. To achieve this goal, we design a graph-based framework capable of representing the structure and evolution of Monero transactions.

Read the complete text