Differentiable Equilibrium Averages: From Zwanzig to Sampler-Agnostic Gradients for MOF Design
A research-paper-grade course that reconstructs the kUPS differentiable-μVT proposal from scratch. You will derive the score-function / Zwanzig identity ∂⟨N⟩/∂θ = −β·Cov(N, ∂U/∂θ) from the grand-ca…
End-to-End Learned Image Compression: From Transform Coding to Neural Codecs
A hands-on, matplotlib-heavy course that builds the mental models behind Ballé et al.'s 'End-to-End Optimized Image Compression' (ICLR 2017) — the paper that launched modern neural image compressio…
From Kevin Bacon to Billion-Scale Vector Search: Building ANN Indices from Scratch
A hands-on systems engineering course that builds a working vector search engine piece by piece. Starting from brute-force search and the curse of dimensionality, you'll implement greedy graph trav…
Self-Compressing Neural Networks: Learning to Quantize from Scratch
Reproduce the Self-Compressing Neural Networks paper (arXiv 2301.13142) from scratch using PyTorch. Build a neural network that learns its own compression during training by making quantization bit…