Ideas from my research and beyond, explained with doodles, illustrations, and animations.
Series
What can a neural network compute at all, before anyone trains it? I illustrate David Chiang's course notes, one topic at a time.
His Theory of Neural Networks (Notre Dame, 2024) covers perceptrons, universal approximation, RNNs and automata, transformers and logic, state-space models.
First piece in progressSeries
An algorithm is a recipe the computer can't improvise on. I animate the classics from Cormen et al., Introduction to Algorithms, one step at a time.
Sorting, search trees, graphs, dynamic programming.
PlannedMy research, told with a coffee machine. You want raspberry flavor, but it keeps tasting like banana. Is it the dials, or the design? I am the mechanic who takes the hood off and finds out.
In progressSeries
Theranos from the court record: trial exhibits, testimony and rulings, put in order by date. Each entry links to the document it comes from.
Have access to records or data from the case? I would love to hear from you: lena.strobl@yale.edu.
Research in progress