Jokhongir Khayrullaev
Position
Consultant, Geophysicist developing inversion algorithms with wave physics, AI, and mathematics
Affiliation
Research groups
Short info
Jokhongir Khayrullaev is a mathematically oriented geophysicist developing methods to image the subsurface and invert rock properties using seismic data. His current research interest lies at the intersection of ray-tracing–based asymptotic wave equation solutions and deep learning approaches.
Work
Recently, Jokhongir Khayrullaev has been working on inverting subsurface P-wave velocities from time-migration velocities. He has developed several methods; as input data for the methods, temporal derivative of the time-migration velocities, Dix slowness is used; the methods leverage kinematic and dynamic seismic ray tracing for simulating the Dix slowness on the subsurface velocities. His research has focused on topics as:
- Deterministic and Bayesian implicit neural representations for velocity inversion
- Bayesian estimation of epistemic and heteroscedastic aleatoric uncertainties
- Field validation: impact of inverted velocities and uncertainties on field data
Outreach
Fluent in English, good in bokmål
Publications
Poster
- Khayrullaev, Jokhongir; Iversen, Einar; Jakobsen, Morten et al. (2025). Estimation of interval P-wave velocities from Dix velocities using a neural-network parameterization. (external link)
- Khayrullaev, Jokhongir; Iversen, Einar; Jakobsen, Morten et al. (2025). Estimation of the P-wave Velocity Model from Dix Velocities Using a Neural Network Parameterization. (external link)