Elliptic, the blockchain analysis provider based in London, has released the largest dataset of labeled Bitcoin (BTC) transactions in the world, termed as Elliptic Data Set, in collaboration with researchers from the MIT (Massachusetts Institute of Technology) and technology giant IBM. This labeling technique will emphasize distinct characteristics in transactions which, in turn, would help identify illicit crypto activities such as money laundering.
The dataset is available publicly and can be utilized by researchers, as well as open-source developers, to spot attributes which are distinct to legit or illicit transactions through the machine learning algorithms. As per the reports, the start-up has formed this dataset by using a few machine learning techniques, such as Random Forest, Multilayer Perceptrons, GCN (Graph Convolutional Networks), and Logistic Regression.
The dataset by Elliptic consists of as many as 200,000 Bitcoin transactions that are worth the value of 6 billion dollars. The researchers identified and labeled transactions that demonstrated criminal or suspicious characteristics so that development, as well as testing of newer predictive methods, can be carried out.
Tom Robinson, the Co-Founder of Elliptic, stated that they have labeled both kinds of transactions based on their research – those performed by illegitimate actors like fraudsters, ransomware operators, and dark marketplaces and the ones carried out by legal actors such as wallet services, regulated exchanges, and merchants to name a few. When new data is applied, the software is capable of extracting transactions which match such patterns, explained Robinson.
The firm’s statement also said that the same methods can be applied to a variety of digital currencies or blockchain-backed assets, right from Facebook’s upcoming crypto Libra to Ethereum. With this endeavor, the company focuses to assist its clientele better detect illegal transactions, reduce compliance cost, and remove criminal activities from the crypto industry.
It’s not an unknown fact that the crypto space witnesses billion dollars’ worth money laundering every year. With such technological advancements, it will become possible to identify bad actors who would make the financial system more trust-worthy and safer. It will also boost the legitimacy of cryptos for governments across the world.
However, some people are raising privacy concerns attached to such technological innovations. Resolving those queries, Elliptic declared that they don’t receive any personally identifiable details, like name or address, in the data they receive from financial service entities and exchanges. Still, the data can be utilized for connecting many transactions to the corresponding client ID to avert financial crimes.
The researchers had also issued a paper titled ‘Anti-Money Laundering in Bitcoin: Experiments with Graph Convolutional Networks for Financial Forensics’ on 5th August at the Knowledge Discovery and Data Mining Conference.