Helio's main feature consists of its search algorithm, which utilizes an optimized graph structure to efficiently store and process connections between data. All user graphs are stored in our database, and are used as data points for training our model.
The idea is that because all data is on-chain, we are able to determine patterns and preemptively act on them. With optimized processing, even CEX data can be associated back to specific wallets, as data patterns are exposable in any scheme.
Our algorithm runs on a heuristic style DFS search, where more promising paths are explored and unpromising paths are discarded. What is considered an unpromising path is continuously iterated upon.
Because of the limits of accessing on chain data, a heuristic style search is necessary to limit the time complexity and augment the speed of the algorithm.
Of course, we can't fully reveal our secret sauce, but just know, this is the worst our algorithm will ever be.
Last updated
