Films · Computer science · Machine learning

Prof. Dr. Dirk Schieborn
Applied Mathematics 3 · Reutlingen University
Lecture film · Applied Mathematics 3
Fog, Kangaroo, Neighbours: How a Model Learns and Decides
Lecture film · Applied Mathematics 3
Fog, Kangaroo, Neighbours: How a Model Learns and Decides

Prof. Dr. Dirk Schieborn
Applied Mathematics 3 · Reutlingen University
We choose the shape; the data set the knobs. Gradient descent finds the setting in the fog, step by step. And in classification: a border should hit the future, not every old point.
heyprof.app/f/am3-nebel-nachbarn-en
More films: Signature · Prof. Dr. Dirk Schieborn
We store questions without any link to you; selected ones appear anonymously as frequent questions.
Fog, Kangaroo, Neighbours: How a Model Learns and Decides
An explainer film in seven chapters, made from a real lecture: a model as a shape with knobs, gradient descent as a ball rolling through the fog into the valley of the smallest error, the learning rate between snail and kangaroo, ski tracks for bias and variance, a bank deciding on loans, a separating plane in space, and k nearest neighbours. Pause anywhere and ask.
How a film works on heyprof
Every film here listens. Pause it, ask it questions and let it continue with your own examples.
- PauseTap pause or press the space bar. The film waits right at that moment.
- AskTap “Ask” and type what you want to know. The answer comes inside the film, with picture and voice.
- Steer itTell it what to show, for example “Do this with z = 2 − 3i”. The film sets up the app and works it out.
- Try it yourselfWhile it is paused, drag points and sliders yourself. “Continue” takes you back into the film.
Without an account: five questions per visit. With an account you can also ask by voice.
Comments