ACK-J MAGIC Grant Final Update!

1M Ago
# Non-Technical tl;dr This security audit of Monero's ring signature resilience to AI-attacks proposes three different models, the best of which achieved 13.3% accuracy predicting the true spend of mainnet transactions. This revealed that blockchain metadata could provide a marginal improvement of 4.3% greater than random guessing. To hopefully inspire future works, the code to collect, process, and train ML/DL models, as well as the datasets used, are freely available on the [Github repo]( The final report, containing all technical details, is also published on the GitHub repo and can be accessed [HERE.]( `If you would like to read the full technical update please refer to the following Github post:` [CLICK HERE!](