Debido al crecimiento de las redes inalámbricas, se hace necesario un uso eficiente del espectro, una solución es Radio Cognitivo. En una de las etapas de esta tecnología se realiza el sensado de espectro, es decir determinar en una frecuencia si existen usuarios primarios y en caso de no existir, ocupar el espectro disponible; esto se logra al aplicar técnicas de sensado, cada técnica requiere de recursos hardware y puede identificar diferentes características de una señal. Hoy en día se utilizan altas frecuencias de propagación, es necesario que la etapa de procesamiento realice un muestreo Sub Nyquist, es decir a menos del doble de la máxima frecuencia, una alternativa de solución es utilizar un algoritmo basado en Matrix Completion, denominado por los autores como IZMA-SD. Los resultados muestran que en diferentes señales muestreadas al %75 de Nyquist y bajo diferentes SNR, al aplicar el algoritmo se realiza la reconstrucción de la señal, a la cual se puede aplicar las técnicas de sensado.

Introducción: La Radio Cognitiva (CR) hace un uso eficiente del recurso radioeléctrico, para ello realiza la Detección de Espectro (SS) con el fin de identificar el espectro disponible. Pero debido a la rápida evolución de los transceptores, la microelectrónica y las altas frecuencias de propagación, se hace necesario que en CR se apliquen algoritmos de SS en bandas de frecuencia y se realice un muestreo inferior a la tasa de Nyquist. Objetivo: Adaptar un algoritmo para Detección de Espectro Sub-Nyquist en Banda Ancha (WBSS) para redes de CR mediante la Compleción de Matrices (MC) que integra valores semilla a partir de las muestras conocidas, con el fin de completar las entradas no muestreadas de la banda a evaluar, reconstruir las señales e identificar el espectro disponible. Metodología: Se realizó una adaptación al algoritmo Aproximación Matricial de la Zona de Interés (IZMA), para ello se diseña la etapa de reconstrucción y se elige un método de detección de espectro en banda estrecha para conformar el banco de detectores; el algoritmo que se denomina IZMA_SV es evaluado a nivel de simulación, por tanto se reconstruyen señales determinísticas en diferentes SNR y se identifica el estado del canal como ocupado o libre. Resultados: Las simulaciones indican que el algoritmo adaptado presenta diferencias entre los valores conocidos de la matriz de muestreo M y la matriz recuperada X en SNR inferiores a -8 dB, mientras que la diferencia tiende a cero en SNR superiores a 2 dB. Conclusiones: El algoritmo IZMA-SV logra reducir el número de operaciones para llegar a la matriz aproximada X, reconstruyendo señales muestreadas al 75% de la tasa Nyquist y aún con un muestreo del 20% se mantienen las características de la señal que hacen posible la detección de espectro en banda ancha.

Introduction: This article stems from the doctoral thesis “Characterization and Classification of Seismo-Volcanic Signals Using Artificial Intelligence Techniques and Data Analytics” at the University of Cauca (2022–2026). Problem: Real-time seismic monitoring requires a rapid and accurate classification process for various types of events to assess volcanic activity and thus activate possible early warnings. However, this field faces a fundamental operational bottleneck: in a newly instrumented volcano, monitoring networks record different types of earthquakes with a lower volume of data than required by standard deep learning classifiers, which need hundreds or thousands of labeled samples to achieve adequate performance. Objective: To overcome the data scarcity problem, we developed and evaluated a Multimodal Prototypical Network (MPN) designed to maintain good classification performance even when labeled data are extremely limited. Methodology: Our approach merges CLIP ViT-B/32 visual embeddings, extracted directly from seismic spectrograms, with a 19-dimensional tabular descriptor vector based on the physical characteristics of the signals under study. This multimodal input is processed using a 5-fold ensemble strategy driven by focal loss. The model was rigorously evaluated using the Chiles-Cerro Negro Volcanic Complex (CCNVC) dataset, categorizing three distinct signal classes: Volcanic-Tectonic (VT), Long-Period (LP), and Very-Long-Period (VLP) events. Results: Under full supervision (6,366 labeled samples), the MPN achieved exceptional performance with a 98.73% balanced accuracy and a 99.89% ROC AUC. Crucially, in a few-shot scenario with only 33 labeled waveforms (14 VT, 14 LP, and 5 VLP), the model retained a highly robust 85.27% balanced accuracy and a 95.36% ROC AUC. Furthermore, physical consistency analysis confirmed that the network’s decisions are driven by spectral morphology (amplitude–geometry correlation, r = 0.248) rather than simple amplitude geometry, aligning with the expected physics of volcanic source mechanisms. Conclusion: These findings establish intermodal transfer from large-scale vision-and-language pretraining as a highly viable and physically consistent strategy for volcanic monitoring environments with a scarcity of labeled data. Originality: Instead of relying solely on raw waveform statistics or standard deep learning, this work uniquely combines foundation models (CLIP) with strict tabular constraints based on physical features to solve the specific few-shot labeling problem inherent in operational volcanology. Limitations: While performance with limited labeled data is noteworthy, the absolute lower bound for labeling requirements (e.g., fewer than 5 VLP samples) was not tested. Furthermore, the model’s physical consistency is tied to the specific origin mechanisms of VT, LP, and VLP events in the CCNVC; applying this exact multimodal framework to signals with greater overlap (such as hybrid events or tremors) might require recalibrating the tabular vector based on the physics of those signals.

Cross-domain transfer of seismo-volcanic classifiers across geologically distinct volcanic systems remains a critical challenge: models trained on one volcano degrade when deployed at a new site without target-domain labels. This work proposes the Phase-Modulated Hypergraph Fusion (PMHF) model, a trimodal deep architecture that jointly exploits CLIP-encoded spectrograms, waveform images, and physics-informed seismic features to discriminate Long-Period (LP) from Volcano-Tectonic (VT) events in an adaptation domain setting. PMHF integrates three novel modules: a Spectral Calibrated Physics Gate (SCPG) that recalibrates tabular features via class-conditional resonance matrices; Dynamic Hypergraph Convolutions (DHGC) that propagate visual features through probability-conditioned hyperedges; and a Trimodal Episodic Cross-Attention (TECA) mechanism that fuses all pairwise modal interactions through persistent memory. Three leave-one-volcano-out scenarios evaluated across Galeras, Cumbal, and Chiles-Cerro Negro (bootstrap N=1000) show that PMHF achieves mean balanced accuracy 0.765, peaking at 0.868±0.009 (MCC =0.743±0.017) for Galeras+Cumbal→Chiles-Cerro Negro, where bimodal baselines plateau below 0.659 and tabular models exhibit systematic class inversion (MCC =−0.405). Domain adaptation diagnostics confirm that trimodal fusion achieves the tightest and most stable source-target feature alignment, enabling robust volcanic monitoring without retraining.

This study presents a comparative evaluation of three Deep Learning (DL) models for the automatic classification of seismo-volcanic events from spectrographic representations. The models considered include an optimized Lightweight Convolutional Neural Network (LCNN) (TFK) and two Transfer Learning (TL) based architectures: RegNetX002 and EfficientNetB0. The evaluation is performed on an unbalanced multiclass dataset including Volcano-Tectonic (VT), Long Period (LP) and Very Long Period (VLP) event signals, using stratified partitions for training, validation and testing. Different Batch Sizes (BS) are analyzed, observing that EfficientNetB0 achieves superior results in metrics such as Accuracy, Recall, F1-Score and Accuracy, while TFK stands out for its high competitiveness with less complexity in its architecture. The learning curves show stable convergence in all three models; EfficientNetB0 and RegNetX002 show minimal validation loss and no overfitting. Although TFK exhibits higher variance, it exceeds 98% accuracy in the validation phase. An adversarial test based on label perturbation is incorporated in order to assess the generalizability of the models. Results indicate that all approaches maintain robust performance even under moderate noise and unbalance conditions, supporting their applicability as efficient solutions for real-time volcano monitoring systems.

Reliable discrimination of Volcano-Tectonic (VT) and Long-Period (LP) seismic signals is central to volcanic surveillance, yet operational deployment is hindered by pronounced inter-volcano domain shift, scarce annotated catalogues, and heterogeneous acquisition instrumentation. We present a self-supervised trimodal deep learning framework that jointly encodes spectrograms, raw one-dimensional waveforms, and physical tabular features, and we evaluate its capacity for cross-domain generaliza-tion across three leave-one-volcano-out transfer settings among Galeras, Cumbal, and Chiles-Cerro Negro (Colombia). Beyond classification accuracy, we systematically characterise the geometry of the learned latent space using class- and domain-separability landscapes projected into two dimensions, quantifying silhouette coefficients, Davies–Bouldin indices, separation indices, and decision-boundary alignment for every ablation variant. The trimodal framework (SSL Trimodal) attains the highest bootstrap balanced accuracy on the Galeras+Cumbal→Chiles-Cerro Negro transfer (0.9080±0.0072) and the most favourable class-separation index (Sep-Idx_2D = 2.06) among all eight evaluated variants in that scenario. It further achieves the best class-separation index (1.72) even on the most divergent transfer setting, Galeras+Chiles-Cerro Negro→Cumbal, despite not attaining the highest accuracy there. This dissociation between point accuracy and embedding quality is, we argue, the more in-formative signal for assessing cross-domain robustness than accuracy alone. We report this result alongside its limitations transparently: the CORAL covariance-alignment baseline outperforms the trimodal framework on two transfer settings with comparable mean accuracies ( 0.81), underscoring that discriminative contrastive pretraining and second-order domain alignment address complemen-tary, rather than redundant, failure modes. To our knowledge, this is the first trimodal self-supervised framework reported for cross-domain generalization in seismo-volcanic signal classification.