Bookmarks
George Hotz | Programming | twitchchess | a simple neural chess AI | Part1
Live coding session where George Hotz designs and trains a simple neural-network chess engine, examining model architecture, training loop, and gameplay integration.
Simple Artificial Neural Network entirely in assembly language
Demonstrates building and training a single-layer neural network entirely in x86-64 assembly language, covering forward pass, MSE loss, back-propagation, and low-level numeric routines.
What Matters for Model Merging at Scale?
Technical summary of a current arXiv paper on large-scale model merging, providing up-to-date insights for ML practitioners.
Dense Associative Memory in Machine Learning
Research talk on Dense Associative Memory networks, exploring high-capacity energy-based models for pattern storage and retrieval.