Learning
Learning PINN Models for cases in between seismic data and well logs data
·1619 words·8 mins
Traditionally, this process involves multiple steps: first, inversion of seismic data to estimate velocities (through methods like acoustic inversion), and then using these velocities to predict petrophysical properties. The main limitation of these methods is that they often do not incorporate uncertainty quantification, leading to suboptimal predictions. Machine learning, and more specifically physics-informed neural networks (PINNs).