Mathematical and computational modeling for seismic image enhancement in emerging Colombian basins using domain adaptation
Vicerrectoría de Investigación y Extensión, Universidad Industrial de Santander
Computational geophysics · Inverse theory · Scientific AI
Ana Gabriela
Mantilla Dulcey
Reconstructing the hidden Earth through wave physics, optimization and machine intelligence.
Geologist · M.Sc. in Geophysics · Scientific Machine Learning Researcher
01Wave physics
02Inverse theory
03Scientific ML
Academic formationM.Sc. Geophysics
Universidad Industrial de Santander
Verified journal record3 articles
2 first-author
Methodological coreInverse problems
Physics-aware learning and optimization
Open research13 repositories
11 with public documentation
Research agenda
Mathematical foundations for scientific inverse problems
How can mathematical structure, physical knowledge, and learned representations be combined to produce inverse methods that are stable, identifiable, uncertainty-aware, and scalable enough for scientific discovery?View the complete research agenda
Research focus / 01
Learning the Earth from waves.
My work sits at the interface of full-waveform inversion, computational imaging and scientific machine learning: designing methods that recover high-resolution subsurface structure while respecting the governing physics.
- Forward model
- Wave propagation
- Inference
- Large-scale optimization
- Prior
- Geology-aware learning
Selected scholarship
Evidence behind the agenda
202620242025
First author · journal article · IEEE Transactions on Geoscience and Remote Sensing
Imaging-Guided Pattern Optimization for Seismic Acquisition Geometries Design With Deep Learning
Open published recordFirst author · journal article · Gondwana Research
Porphyry-type mineral prospectivity mapping with imbalanced data via prior geological transfer learning
Open published recordjournal article · The Leading Edge
Poststack seismic data denoising via dynamic guided learning
Open published recordResearch practice
From theory to field-scale inference
New computational technologies for the joint processing and inversion of gravity, magnetic, and magnetotelluric data using physics-guided deep learning for multicriteria characterization
MINCIENCIAS-ANH
Reconciliation of land-use practices and ecosystem sustainability through agrogeophysics, artificial intelligence, and citizen science
Vicerrectoría de Investigación y Extensión, Universidad Industrial de Santander
Scientific breadth
Experience spans seismic imaging, gravity-magnetic-magnetotelluric inversion, geothermal 3D modeling, SAR interpretation, geospatial AI, and physics-guided learning.
View the full trajectoryNews
Recent recognition
FIRENET places second nationally in the Datos a la U competition
Ana Gabriela Mantilla Dulcey and teammates León Santiago Suárez Rodríguez and Luis Miguel Rodríguez López were recognized for a geospatial AI platform supporting forest-fire management.