AI RESEARCH
ZKBoost: Zero-Knowledge Verifiable Training for XGBoost
arXiv CS.LG
•
ArXi:2602.04113v3 Announce Type: replace-cross Gradient boosted decision trees, particularly XGBoost, are among the most effective methods for tabular data. As deployment in sensitive settings increases, cryptographic guarantees of model integrity become essential. We present ZKBoost, the first zero-knowledge proof of