This talk presents interim results from a nondestructive tunnel inspection study initiated in February 2026. Using topological time-series analysis, we demonstrate that:

• - High Efficiency: Consistently high accuracy is maintained even with limited datasets.
• - Linear Separability: Defect detection is clearly achieved using a simple linear SVM, proving the robustness of the topological features.
• - Physical Consistency: t-SNE visualization confirms that non- defective states cluster together, while defective states are distinctly separated and show clear depth-dependent variations.

Time permitting, I will also briefly introduce an NMF-based analysis of African Swine Fever to discuss mathematical modeling of vaccine efficacy.