M3SpaDE (Multi-Modal Model for predicting Spatial Drug Efficacy) is a versatile computational framework designed for predicting drug sensitivity in spatial transcriptomics data. It is resolution-agnostic, capable of processing data ranging from single-cell to spot-level resolutions, and supports generalizable prediction of responses to previously unseen drugs based on their chemical structures. M3SpaDE enables the following tasks: Binarized Sensitivity PredictionPerforms binary classification of drug sensitivity at the single-cell or spot level (Sensitive vs. Resistant). Spatial Autocorrelation AnalysisQuantifies global spatial dependency and clustering patterns using Join Count statistics. Combinatorial Therapy AssessmentPredicts and evaluates drug sensitivity outcomes for drug combinations.
These are multimodal dataset objects and trained model parameters used in the study. The files are organized in pairs, where each multimodal dataset (.h5mu file) corresponds to a trained model parameter file (.pt) generated using the MIMA (Multimodal Integration with Modality-agn…
Sensor Validator v3.5 is a Python-based framework for adaptive validation of environmental and chemical sensor systems. The platform combines multi-modal feature extraction, anomaly detection, machine-learning classification, drift monitoring, automatic recalibration, hardware ab…
Abstract: Deepfake technology, driven by generative models such as GANs and diffusion architectures, has enabled the creation of highly realistic manipulated media capable of deceiving both visual and auditory perception. Such forgeries pose significant risks to identity verifica…
Microscopic deformation stage recognition from molecular dynamics (MD) trajectories is crucial for understanding the evolution of material damage; however, traditional empirical analysis and black-box single deep learning models lack both high-throughput spatiotemporal modeling a…
A hybrid deep learning framework for monthly precipitation prediction in mountainous areas of Boyacá, Colombia. Combines Graph Neural Networks (GNN) with temporal attention mechanisms and ConvLSTM architectures for accurate spatiotemporal forecasting. This implementation includes…
Surveillance anomaly detection systems built around a single monolithic deep network are difficult to interpret, brittle to distribution shift, and offer operators no rationale on which to act. This paper presents VisionGuard, an explainable deep learning framework that reorganiz…