New benchmark evaluates autoregressive models for mineral exploration drillholes

Researchers have introduced DrillBench, a new benchmark designed to evaluate autoregressive models for mineral exploration drillhole data. The benchmark, comprising 49,671 drillholes from Western Australia, focuses on predicting deeper geological strata based on sequential data. Experiments with classical, geostatistical, and neural models revealed that while spatial conditioning offers local benefits, autoregressive models trained on lithology sequences demonstrate more robust generalization across geological shifts. AI

IMPACT This benchmark could advance the application of autoregressive models in geological surveying and resource exploration.

RANK_REASON The item is a research paper introducing a new benchmark for evaluating autoregressive models in a specific domain. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• alphaXiv
• CatalyzeX
• Connected Papers
• CORE Recommender
• DagsHub
• DrillBench
• Gotit.pub
• Hugging Face
• IArxiv Recommender
• Influence Flower
• Litmaps
• ScienceCast
• scite Smart Citations
• Western Australia

• paper
• other

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