SPAMoE: New Deep Learning Framework Improves Full-Waveform Inversion for Subsurface Imaging
Researchers have proposed SPAMoE, a spectrum-aware deep learning framework for full-waveform inversion (FWI) that reconstructs high-resolution subsurface velocity models more accurately than existing methods. The approach addresses a core limitation of prior neural network methods—frequency entanglement of multi-scale geological features—by combining a spectral-preserving encoder with a dynamic Mixture-of-Experts routing system. On ten benchmark datasets, SPAMoE reduced mean absolute error by 44.4% relative to the best previously reported baseline, potentially advancing geophysical imaging for applications such as oil exploration and earthquake monitoring.
Full-waveform inversion is a computationally demanding and mathematically ill-posed technique used to reconstruct detailed subsurface velocity models from seismic wave data, with applications in resource exploration and geohazard assessment. Existing deep learning approaches, including convolutional neural networks and single-paradigm neural operators, have struggled with frequency entanglement—the inability to separately model geological features that operate at different spatial scales. SPAMoE addresses this by introducing a Spectral-Preserving DINO Encoder that enforces a minimum ratio of high-to-low frequency energy in learned representations, preventing the collapse of fine-grained detail. A Spectral Decomposition and Routing mechanism then dynamically assigns different frequency bands to a Mixture-of-Experts ensemble composed of three distinct neural operator architectures: FNO, MNO, and LNO. Evaluated across all ten sub-datasets of the OpenFWI benchmark, the framework achieves a 44.4% average reduction in mean absolute error compared to the best officially reported baselines. The authors have released their code and data publicly, facilitating independent replication and further development. The work is currently a preprint on arXiv and has not yet undergone formal peer review.
What's missing
As a preprint, SPAMoE has not undergone formal peer review, and independent replication on real-world seismic datasets (beyond the synthetic OpenFWI benchmarks) has not been reported. Generalization to noise levels and acquisition geometries outside the benchmark conditions remains an open question.
What different sources said
- arXiv cs.LGCenter
SPAMoE: Spectrum-Aware Hybrid Operator Framework for Full-Waveform Inversion
Related
Gut Bacteria Enzyme Found to Break Down Heat-Processed Food Compounds, Producing Novel Biogenic Amines
Researchers have discovered that an enzyme in common gut bacteria can degrade N-epsilon-carboxymethyllysine (CML), a compound formed during thermal food processing, producing previously unknown biogenic amines. The enzyme, ornithine decarboxylase SpeC from enterobacteria, acts on CML and related modified lysine derivatives through a low-level 'underground' catalytic activity. This finding suggests a previously unrecognized communication axis between thermally processed dietary compounds and gut microbial physiology, with potential implications for host health.
Full-Length Gene Sequencing Reveals Two Distinct Bacterial Communities in Black-Legged Ticks Expanding Into Canada
Researchers used Oxford Nanopore full-length 16S rRNA gene sequencing to characterize the microbiome of Ixodes scapularis black-legged ticks collected in Nova Scotia, Canada, distinguishing between tick-adapted bacteria and environmentally acquired bacteria. The study comes as I. scapularis — the primary vector of Lyme disease — is rapidly expanding northward into Canada due to climate change. The findings suggest that environmentally derived bacteria in tick microbiomes are not mere contamination, which has implications for how tick microbiome data is collected and interpreted across surveillance studies.
Study Identifies Metabolic Link Between Cell Envelope Stress and Biofilm Formation in Bacteria
Researchers have discovered that the metabolite acetyl-CoA directly inhibits enzymes that degrade the bacterial signaling molecule c-di-GMP, connecting cell envelope biosynthesis stress to biofilm formation in Pseudomonas aeruginosa. The study found that sub-inhibitory concentrations of antibiotics targeting early peptidoglycan biosynthesis — but not other antibiotic classes — elevate c-di-GMP levels by reducing phosphodiesterase activity, with acetyl-CoA competing for the enzyme active site. Because the relevant enzyme domain is broadly conserved across bacterial species, this checkpoint mechanism may be widespread and could have implications for understanding antibiotic-induced biofilm responses.