Finding Illicit Crypto Deals by Tracking Network Shortest Paths

Tracking down illegal transactions in crypto can feel like trying to find a needle in a massive digital haystack. I recently came across a study that looks at how investigators can zoom in on bad actors without pulling in tons of useless data. Blockchain transaction networks are huge, so pulling out just the shady parts is a real challenge.

The researchers wanted to improve how investigators pull out illicit transaction networks using an algorithm called SPNSA. This algorithm connects the dots by finding the shortest paths between a small starting list of entities, but its results depend heavily on which entities you pick first. To fix this problem, the team tested different ways to select starting points and introduced a new motif-based selection method. They put their idea to the test on the Elliptic Bitcoin transaction graph along with the IBM-AML benchmark datasets named HI-Small, LI-Small and HI-Medium. Their strategy pulled out subgraphs with a much higher ratio of illegal activity compared to random or degree-based choices on Elliptic and the IBM-AML HI variants. Best of all, it managed to highlight those bad transactions while keeping the final output compact.

Curious about the details? Look up Extracting Illicit-Rich Subgraph from Transaction Networks via Feed-Guided Shortest Paths at arxiv.org/abs/2610.13280. The authors are Doniyor Koshimbetov, E. Şule Yazıcı and Asha Rao.

This is not financial advice. The information provided is for educational and informational purposes only.

The cover image was generated by AI.