Transformer Encoders: Or, How Multi-Head Attention Listens to Everything Simultaneously Without Missing the Point
[This is Part 3. Part 1 covers autoencoders. Part 2 covers self-attention.]
Autoencoders: Or, How to Compress Everything You Know Into Two Numbers and Still Recognize Your Cousin
Mathityahu had a cousin — Devorah — who kept every photograph the family had ever taken. Shoeboxes. Closets. An entire drawer dedicated solely to blurry Bar Mitzvah photos from 1994. Her apartment was one fire hazard away from a documentary.
Self-Attention: Or, How to Listen to Everyone at Once Without Going Completely Meshuggeneh
[This is Part 2. Part 1 — The Forgetful Scribe covers autoencoders.]
in Data Science
t-SNE and UMAP: The Methods Your Statistics Professor Warned You About
In Part 1 we tried to bring order to chaos using PCA and SVD.
in Data Science
PCA and SVD: Or, How to Stop Drowning in Data and Start Complaining About Less of It
My bubbe kept every piece of information she had ever received. Every receipt, every letter, every grudge — all of it, organised in a system that made perfect sense to her and to nobody else on earth. You’d ask her one simple question and she’d hand you fourteen folders and a story about something your uncle did in 1987.