Enhancing ground modeling through artificial intelligence

(2026)

Files

Sagdic_55601900_2026.pdf
  • Closed access
  • Adobe PDF
  • 11.77 MB

Details

Supervisors
Faculty
Degree label
Abstract
The thesis "Enhancing Ground Modeling Through Artificial Intelligence" investigates the use of sequential deep learning, specifically LSTMs, BiLSTMs and Transformers, to create more geologically accurate 3D ground models by integrating sparse borehole data with continuous geophysical surveys. By treating soil stratigraphy as a sequence rather than independent data points, the study aims to overcome the noisy, "speckled" predictions typical of traditional machine learning models like CatBoost.