Researchers have introduced DeXposure-FM, a novel time-series graph foundation model designed to quantify and forecast credit exposure within decentralized finance (DeFi) networks. This model aims to address the contagion risks inherent in DeFi's interconnected token system, which can lead to uncontrolled spread of financial shocks. DeXposure-FM utilizes a graph-tabular encoder and has been trained on a substantial dataset comprising over 43.7 million entries across thousands of protocols and blockchains. The model demonstrates superior performance on machine learning benchmarks compared to existing graph foundation models and temporal graph neural networks, and it provides tools for macroprudential monitoring and stress testing in DeFi. AI
IMPACT This model could enhance the stability and risk management of decentralized financial networks by providing better tools for forecasting and monitoring credit exposures.
RANK_REASON This is a research paper describing a new model and dataset. [lever_c_demoted from research: ic=1 ai=1.0]
- alphaXiv
- arXiv
- CatalyzeX
- DagsHub
- decentralized finance
- DeXposure-FM
- Fengxiang He
- Gotit.pub
- Hugging Face
- ScienceCast
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
