An agent using XGBoost Cox survival models and counterfactual simulation prevents Aave v3 liquidations where static health-factor rules fail, achieving zero worsening rate on 4,882 high-risk profiles while ignoring dust events.
The Predictability of Risk: Linking User Action Frequency to Survival Model Performance
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
From Risk to Rescue: An Agentic Survival Analysis Framework for Liquidation Prevention
An agent using XGBoost Cox survival models and counterfactual simulation prevents Aave v3 liquidations where static health-factor rules fail, achieving zero worsening rate on 4,882 high-risk profiles while ignoring dust events.