Beyond Photons: Passive Acoustic Sensing for Autonomous Vehicles
Exploring PAMVON, a passive acoustic monitoring system for emergency vehicle detection, and the challenges preventing its production adoption.
Read more
Locality, Learning, and the FFT: Why CNNs Avoid the Fourier Domain
A unified treatment of 1D signal convolution, 2D image convolution via the convolution theorem, and graph convolution as a spectral operation on the normalized Laplacian.
Read more
Hot & Cold Spectral GCNs Part 2: Heat Diffusion, Neural Networks, and Cold-Start Recommendations
Applying Spectral GCNs to heat diffusion on a 3-D torus: the heat kernel filter, PyTorch implementation, Chebyshev approximation, and the cold-start connection.
Read more
Hot & Cold Spectral GCNs Part 1: From the Fourier Transform to Spectral Graph Convolutions
How the classical Fourier Transform generalises to arbitrary graphs: deriving the Graph Fourier Transform from the Laplacian eigenbasis and building up to the one-layer Spectral GCN.
Read moreMatching MATLAB's resample function in Python
A drop-in Python implementation of MATLAB's resample function using scipy's polyphase filter — matching MATLAB's output exactly, with no signal-processing background required.
Read moreRecents
