معرفی
Malcolm Sambridge is a researcher at the Research School of Earth Sciences, Australian National University. His work focuses on advancing Bayesian inference methods for geophysical problems, particularly through the development of trans-conceptual sampling frameworks.
In the article "Trans-Conceptual Sampling: Bayesian Inference With Competing Assumptions", Sambridge and colleagues propose a novel approach to Bayesian sampling that generalizes trans-dimensional methods, enabling exploration across models with varying conceptual assumptions, basis functions, and forward physics. This framework eliminates the need for bespoke parameter transformations, allowing automated sampling with modern algorithms like Hamiltonian MCMC.
The research demonstrates that trans-conceptual sampling achieves results comparable to reversible-jump MCMC in standard cases but extends applicability to scenarios where traditional trans-dimensional methods would fail, such as models with non-local basis functions or heterogeneous parameterizations. Applications range from airborne electromagnetic inversion to synthetic seismic tomography, showcasing the method's versatility in handling complex geophysical problems.
