Banca de QUALIFICAÇÃO: LUCAS TORRES MARQUES

Uma banca de QUALIFICAÇÃO de MESTRADO foi cadastrada pelo programa.
STUDENT : LUCAS TORRES MARQUES
DATE: 11/07/2025
TIME: 14:00
LOCAL: Remota via Google Meet
TITLE:

Physics-Informed Kolmogorov-Arnold Networks Approach for Seismic Ray Tracing

KEY WORDS:

Seismic Ray Tracing, Kolmogorov-Arnold Networks, Physics-Informed Neural Networks, Computational Geophysics, Machine Learning.

 

PAGES: 78
BIG AREA: Engenharias
AREA: Engenharia Elétrica
SUMMARY:

The modeling of seismic wave propagation is fundamental for seismic imaging and reservoir characterization in the oil and gas industry. Seismic ray tracing is an efficient technique, but traditional methods face difficulties in complex geological models due to instability and high computational cost. This work proposes the use of Kolmogorov-Arnold Networks with Physics-Informed Learning (PIKANs) to overcome these limitations. The methodology includes smoothing the velocity model with B-spline filtering and using adaptive weights in the physical loss function to improve model training. Two experiments were conducted: the first with a parametrized velocity model and the second with the non-parametrized Marmousi model. The results demonstrate the consistency of the approach, with R2 values close to 1 for the first experiment and above 0.90 for the second.

COMMITTEE MEMBERS:
Presidente - 3216921 - TIAGO TAVARES LEITE BARROS
Interno - 2885532 - IVANOVITCH MEDEIROS DANTAS DA SILVA
Interno - 1153006 - LUIZ AFFONSO HENDERSON GUEDES DE OLIVEIRA
Externo à Instituição - RAMON CRISTIAN FERNANDES ARAUJO