Elliptic Releases Data On $6bn Of Bitcoin Transactions
The data set is meant to help other companies to test their own technologies.
Share this article
If you’re trying to launder money, don’t use Bitcoin. Elliptic, a British company providing blockchain monitoring solutions for regulatory compliance, has teamed up with researchers from the MIT-IBM Watson AI Lab to create a public dataset of legal and illegal Bitcoin transactions. Because each bitcoin fragment can be tracked through the blockchain, the new dataset will allow almost anyone to identify coins which came from illicit sources.
The Elliptic Data Set covers $6 billion worth of Bitcoin transactions, with 4545 transactions (2%) labeled as illicit, defined for their relation to scams, malware, terrorist organizations, darknet markets, or other unlawful activity. Structured as a graph, the dataset contains 203,679 nodes that represent single transactions, connected by edges that seek to represent the flow of bitcoins from one transaction to the other.
Another 42,019 (21%) of the transactions are labeled as legitimate, corresponding to miners, exchanges, wallet providers and other services, while the remaining 77% of transactions come from unknown sources.
The data set has been made public by Elliptic to motivate other actors in the field to develop and test new techniques, allowing them train their own predictive technologies on it.
Elliptic is a provider of blockchain monitoring solutions for regulatory compliance and risk management by cryptocurrency businesses and financial institutions worldwide. The company is pioneering new methods in deep learning for graph or network-structured data, named Graph Convolutional Networks, which are being developed to identify complex money laundering schemes in cryptocurrency.
The data set release coincides with a new paper on the technique that Elliptic scientists have co-authored with researchers from the MIT-IBM Watson AI Lab. The paper, entitled, “Anti-Money Laundering in Bitcoin: Experiments with Graph Convolutional Networks for Financial Forensics,” will be presented by IBM Research Staff Members Mark Weber and Giacomo Domeniconi at the Anomaly Detection in Finance workshop of the Knowledge Discovery and Data Mining Conference (KDD) on August 5, 2019.
Elliptic uses a variety of techniques to discover financial crime on blockchain, and the work with GCNs is part of a wider push for ever-improving identification techniques, according to Chief Scientist and co-founder of Elliptic, Tom Robinson.
“Elliptic uses a range of advanced techniques, including machine learning, to facilitate financial crime detection in cryptocurrencies,” Robinson said. “Our work with researchers from the MIT-IBM Watson AI Lab builds on this, to ensure that our clients have access to the most accurate and effective insights available, reducing their compliance costs and ensuring that their services are not exploited by criminals.”