Convolutional Template Optimization for Noisy Signal Alignment

This project analyzes gravitational wave signals from colliding black holes using shift-invariant correlation landscapes.

A convolutional optimization approach explores the embedding space to localize waveform matches under extreme noise. The heatmaps below visualize confidence regions with annotations marking the ground truth and peak correlation.

This work was conducted under the supervision of Professor John Wright at Columbia University. For the complete implementation, see the GitHub repository and the Project Report.