# Benchmarking Temporal Web3 Intelligence: Lessons from the FinSurvival 2025 Challenge

arXiv: https://arxiv.org/abs/2602.23159 · HTML: https://mlsanity.com/paper/2602.23159/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 12 of its 51 references that are in the map.
- First posted on arXiv 26 Feb 2026.

## Builds on (most cited first, 12 of 12)

- [Optuna: A Next-generation Hyperparameter Optimization Framework](https://mlsanity.com/paper/1907.10902/) (2019)
- [DeepSurv: Personalized Treatment Recommender System Using A Cox Proportional Hazards Deep Neural Network](https://mlsanity.com/paper/1606.00931/) (2016)
- [Codabench: Flexible, Easy-to-Use and Reproducible Benchmarking Platform](https://mlsanity.com/paper/2110.05802/) (2021)
- [Survival regression with accelerated failure time model in XGBoost](https://mlsanity.com/paper/2006.04920/) (2020)
- [SoK: Decentralized Finance (DeFi)](https://mlsanity.com/paper/2101.08778/) (2021)
- [Improved Price Oracles: Constant Function Market Makers](https://mlsanity.com/paper/2003.10001/) (2020)
- [AI-powered Fraud Detection in Decentralized Finance: A Project Life Cycle Perspective](https://mlsanity.com/paper/2308.15992/) (2023)
- [CeFi vs. DeFi -- Comparing Centralized to Decentralized Finance](https://mlsanity.com/paper/2106.08157/) (2021)
- [Deciphering Crypto Twitter](https://mlsanity.com/paper/2403.06036/) (2024)
- [Significant Digits: Using Large-Scale Blockchain Data to Predict Fraudulent Addresses](https://mlsanity.com/paper/2301.01809/) (2023)
- [Smart Contracts, Smarter Payments: Innovating Cross Border Payments and Reporting Transactions](https://mlsanity.com/paper/2407.19283/) (2024)
- [Enhancing Web Spam Detection through a Blockchain-Enabled Crowdsourcing Mechanism](https://mlsanity.com/paper/2410.00860/) (2024)

## Built on by (most cited first, 1 of 1)

- [From Risk to Rescue: An Agentic Survival Analysis Framework for Liquidation Prevention](https://mlsanity.com/paper/2604.14583/) (2026)

## Abstract (arXiv)

Temporal Web analytics increasingly relies on large-scale, longitudinal data to understand how users, content, and systems evolve over time. A rapidly growing frontier is the Temporal Web3: decentralized platforms whose behavior is recorded as immutable, time-stamped event streams. Despite the richness of this data, the field lacks shared, reproducible benchmarks that capture real-world temporal dynamics, specifically censoring and non-stationarity, across extended horizons. This absence slows methodological progress and limits the transfer of techniques between Web3 and broader Web domains. In this paper, we present the textitFinSurvival Challenge 2025 as a case study in benchmarking temporal Web3 intelligence. Using 21.8 million transaction records from the Aave v3 protocol, the challenge operationalized 16 survival prediction tasks to model user behavior transitions.We detail the benchmark design and the winning solutions, highlighting how domain-aware temporal feature construction significantly outperformed generic modeling approaches. Furthermore, we distill lessons for next-generation temporal benchmarks, arguing that Web3 systems provide a high-fidelity sandbox for studying temporal challenges, such as churn, risk, and evolution that are fundamental to the wider Web.