Improving anti-money laundering in bitcoin using evolving graph convolutions and deep neural decision forest
| Date | 28 November 2022 |
| Pages | 313-329 |
| DOI | https://doi.org/10.1108/DTA-06-2021-0167 |
| Published date | 28 November 2022 |
| Author | Anuraj Mohan,Karthika P.V.,Parvathi Sankar,K. Maya Manohar,Amala Peter |
Improving anti-money laundering
in bitcoin using evolving graph
convolutions and deep neural
decision forest
Anuraj Mohan, Karthika P.V., Parvathi Sankar,
Maya Manohar K.and Amala Peter
Department of Computer Science and Engineering, NSS College of Engineering,
Palakkad, India
Abstract
Purpose –Money laundering is the process of concealing unlawfully obtained funds by presenting them as
coming from a legitimate source. Criminals use crypto money laundering to hide the illicit origin of funds
using a variety of methods. The most simplified form of bitcoin money laundering leans hard on the fact that
transactions made in cryptocurrencies are pseudonymous, but open data gives more power to investigators
and enables the crowdsourcing of forensic analysis. With the motive to curb these illegal activities, there
exist various rules, policies and technologies collectively known as anti-money laundering (AML) tools.
When properly implemented, AML restrictions reduce the negative effects of illegal economic activity while
also promoting financial market integrity and stability, but these bear high costs for institutions. The
purpose of this work is to motivate the opportunity to reconcile the cause of safety with that of financial
inclusion, bearing in mind the limitations of the available data. The authors use the Elliptic dataset; to the
best of the authors’knowledge, this is the largest labelled transaction dataset publicly available in any
cryptocurrency.
Design/methodology/approach –AML in bitcoin can be modelled as a node classification task in
dynamic networks. In this work, graph convolutional decision forest will be introduced, which combines
the potentialities of evolving graph convolutional network and deep neural decision forest (DNDF). This
model will be used to classify the unknown transactions in the Elliptic dataset. Additionally, the application
of knowledge distillation (KD) over the proposed approach gives finest results compared to all the other
experimented techniques.
Findings –The importance of utilising a concatenation between dynamic graph learning and ensemble
feature learning is demonstrated in this work. The results show the superiority of the proposed model to
classify the illicit transactions in the Elliptic dataset. Experiments also show that the results can be further
improved when the system is fine-tuned using a KD framework.
Originality/value –Existing works used either ensemble learning or dynamic graph learning to tackle the
problem of AML in bitcoin. The proposed model provides a novel view to combine the power of random
forest with dynamic graph learning methods. Furthermore, the work also demonstrates the advantage of KD
in improving the performance of the whole system.
Keywords Graph deep learning, Knowledge distillation, Graph convolutional networks, Anti-money
laundering, Deep neural decision forest, Anomaly detection
Paper type Research paper
1. Introduction
Bitcoin is a decentralised, pseudonymous, peer-to-peer virtual currency system that
operates purely by an algorithm. Transacting parties can send bitcoin directly without
the need for intermediaries, and the network nodes use cryptography to verify these bitcoin
transactions, which are then recorded in a public distributed ledger called the blockchain.
Hashes serve as the link in the blockchain, i.e. each block includes the previous block’s
unique hash. To modify an entry in the ledger retroactively, a new hash has to be calculated
not only for the block it is in but also for every subsequent block. This requires powerful
ThecurrentissueandfulltextarchiveofthisjournalisavailableonEmeraldInsightat:
https://www.emerald.com/insight/2514-9288.htm
313
Received 27 June 2021
Revised 18 September 2021
10 December 2021
6 March 2022
20 April 2022
5 August 2022
Accepted 14 September 2022
Data Technologies and
Applications
Vol. 57 No. 3, 2023
pp. 313-329
© Emerald Publishing Limited
2514-9288
DOI 10.1108/DTA-06-2021-0167
Improving
anti-money
laundering in
bitcoin
computers because any block that is added will conflict with the existing ones and the other
nodes will automatically reject the alterations. Hence, the blockchain is tamperproof, and
this functionality makes it immune to the risk of losing the data to hackers who look for
vulnerabilities in the system to steal bitcoins. Bitcoin has grown in popularity in recent
decades, owing to its inclusive nature and anonymity. This means that both legitimate and
criminal users can use the system to send money at near-instantaneous speed for little or no
cost, with low barriers to entry, while remaining virtually anonymous without the need for
a public paper trail. Due to this openness, virtual currency systems like bitcoin are
becoming a breeding ground for money laundering activities.
Money laundering using bitcoin creates a negative impact on society due to many
reasons. The most cataclysmic one is corruption (Fauzi et al., 2020). Of the three steps
including placement, layering and integration of money laundering, placement is the
vulnerable step in which the launderers can place the dirty money into a legitimate
system. This is a cause for the decline of economic systems and societal culture as people
are directed towards illegitimate activities. The anonymity provided by the cryptocurrency
cannot be ignored in the discussion of its pessimistic impact on society. Drug dealing and
trafficking are other activities with a stultifying effect on society. The humongous number
of users of bitcoin is advantageous for criminals to carry out illegal activities, and the
reputation of bitcoin is under threat by these wicked activities. The illegitimate use of
bitcoin may supplant the legitimate use and shatter the entire system. Low-income people
are highly inflicted with the consequences of an unsafe financial system. Criminals
generally target the emerging financial sectors for their laundering activities, and these
fluctuate demand, income and growth rates. The socio-economic sector of a developing
country is more likely to be blemished by money laundering due to the limited capacity of
the financial systems. Several organisations that face money laundering are limited in their
access to world markets. The companies that are controlled by launderers or criminals
blend illegitimate and legitimate funds and thus unfairly make profit. All these contribute
to the deprivation of national income and debilitate a country’sfinancial system. The risks
associated with financial institutions pave the way for these systems’integrity and
stakeholder’s trust to deteriorate. The illegal income earned by criminals poses a threat to
an economy, and economies are restricted when they are to be accentuated. Smuggling and
other similar activities are also among the adversities of money laundering. General
solutions to money laundering can be aided by powerful anti-money laundering (AML)
regulations that are performed by financial institutions to achieve compliance with legal
requirements to actively monitor and report suspicious activities. However, the existing
solutions are still not impeccable. There is a need to polish these regulations or to detect new
ones to shield our financial systems and ensure development of the economy.
Elliptic, a cryptocurrency intelligence company focused on safeguarding cryptocurrency
ecosystems from criminal activity, has published a dataset which, to the best of the authors’
knowledge, is the world’s largest labelled transaction dataset publicly available in any
cryptocurrency (Elliptic, 2019). This is a promising step towards AML, and it has motivated
researchers to work on this societally important challenge. In the bitcoin data provided by
Elliptic, the graph is constructed with nodes representing transactions, which are transfers
from one bitcoin address (e.g. a person or a firm) to another. Each transaction consumes the
output of previous transactions and produces outputs that can be used by subsequent
transactions. The graph’s edges reflect the flow of bitcoins from one transaction to the next.
The dataset has 49 graphs sampled from the bitcoin blockchain at various points in time,
each of which is a directed acyclic graph that starts with one transaction and includes
subsequent connected transactions on the blockchain, encompassing about 2 weeks of data.
Among the 203,769 transactions, 21 per cent are labelled as licit (exchanges, wallet
providers, miners and licit services) and 2 per cent as illicit (scams, malware, terrorist
DTA
57,3
314
Get this document and AI-powered insights with a free trial of vLex and Vincent AI
Get Started for FreeStart Your Free Trial of vLex and Vincent AI, Your Precision-Engineered Legal Assistant
-
Access comprehensive legal content with no limitations across vLex's unparalleled global legal database
-
Build stronger arguments with verified citations and CERT citator that tracks case history and precedential strength
-
Transform your legal research from hours to minutes with Vincent AI's intelligent search and analysis capabilities
-
Elevate your practice by focusing your expertise where it matters most while Vincent handles the heavy lifting
Start Your Free Trial of vLex and Vincent AI, Your Precision-Engineered Legal Assistant
-
Access comprehensive legal content with no limitations across vLex's unparalleled global legal database
-
Build stronger arguments with verified citations and CERT citator that tracks case history and precedential strength
-
Transform your legal research from hours to minutes with Vincent AI's intelligent search and analysis capabilities
-
Elevate your practice by focusing your expertise where it matters most while Vincent handles the heavy lifting
Start Your Free Trial of vLex and Vincent AI, Your Precision-Engineered Legal Assistant
-
Access comprehensive legal content with no limitations across vLex's unparalleled global legal database
-
Build stronger arguments with verified citations and CERT citator that tracks case history and precedential strength
-
Transform your legal research from hours to minutes with Vincent AI's intelligent search and analysis capabilities
-
Elevate your practice by focusing your expertise where it matters most while Vincent handles the heavy lifting
Start Your Free Trial of vLex and Vincent AI, Your Precision-Engineered Legal Assistant
-
Access comprehensive legal content with no limitations across vLex's unparalleled global legal database
-
Build stronger arguments with verified citations and CERT citator that tracks case history and precedential strength
-
Transform your legal research from hours to minutes with Vincent AI's intelligent search and analysis capabilities
-
Elevate your practice by focusing your expertise where it matters most while Vincent handles the heavy lifting
Start Your Free Trial of vLex and Vincent AI, Your Precision-Engineered Legal Assistant
-
Access comprehensive legal content with no limitations across vLex's unparalleled global legal database
-
Build stronger arguments with verified citations and CERT citator that tracks case history and precedential strength
-
Transform your legal research from hours to minutes with Vincent AI's intelligent search and analysis capabilities
-
Elevate your practice by focusing your expertise where it matters most while Vincent handles the heavy lifting