New CoGameReg framework improves seismic reservoir property prediction with sparse data

Researchers have developed CoGameReg, a novel semi-supervised regression framework designed to improve the estimation of reservoir properties from seismic data, particularly when labeled well data is scarce. The framework utilizes a cooperative game-based approach to generate shared pseudo-labels by weighting the predictions of various regressors. This method explicitly defines each regressor's contribution to the learning targets for unlabeled seismic samples, enhancing accuracy under limited well supervision. CoGameReg has demonstrated superior performance on public datasets, synthetic data, and a field application, showing significant advantages in predicting reservoir properties with minimal labeled data. AI

IMPACT This framework could enhance the accuracy of subsurface geological models, impacting resource exploration and extraction.

RANK_REASON The item is a research paper detailing a new methodology for seismic reservoir property prediction. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• alphaXiv
• arXiv
• CatalyzeX
• CoGameReg
• DagsHub
• Gotit.pub
• Hugging Face
• IArxiv
• ScienceCast
• Tri-Game Shapley

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →