Learning relaxation time distributions from spectral induced polarization data with a complex-valued variational autoencoder
Spectral induced polarization (SIP) is a geophysical method used to characterize subsurface materials. It measures the frequency-dependent complex electrical resistivity of rocks and soils through the application of an alternating current. Debye and Warburg decompositions are standard methods for analyzing and interpreting SIP data, as they allow estimation of the relaxation time distribution (RTD) of geomaterials. However, conventional decomposition approaches treat measurements independently, work in real-valued spaces despite the complex-valued nature of SIP data, and provide limited uncertainty quantification. These limitations reduce the effectiveness of conventional decomposition on heterogeneous datasets. We reformulate the decomposition approach as an unsupervised machine learning problem and introduce a conditional variational autoencoder (CVAE) that learns a shared mapping from SIP datasets to continuous RTDs. The model is validated on a dataset comprising 140 laboratory and field SIP measurements of granular mixtures, sedimentary and igneous rock, and cementitious materials. The CVAE operates in complex-valued data space and achieves reconstruction errors of 0.53 % and 0.27 % for the imaginary and phase components of resistivity, respectively, with evidence of improvement over conventional methods at corresponding p-values of 0.007 and 0.06. The inferred RTDs are stable and physically consistent, and their total chargeability and mean relaxation time correlate with polarizable grain content and grain size, respectively, with coefficients of determination up to 0.98. An additional contribution of the proposed method is the learned latent representation, which organizes SIP spectra into a structured space. Unsupervised clustering in a two-dimensional projection of this space improves the silhouette score by nearly a factor of two relative to conventional RTD parameters.