Opening Structure of Neural Networks for Deep Learning
Structure of Neural Networks for Deep Learning
- Length
- 53 min
- Level
- Undergraduate
- Recorded
- MIT2019
Chapter 1 of 8: Before the lecture
Strang counts the flat pieces that folds cut in dimensions, reaching binomial coefficients and a recursion.
Deep neural networks build a learning function from training data, and Strang sets up that function for binary classification of feature vectors. After a demonstration on the TensorFlow Playground, he constructs one layer from a weight matrix, a bias offset and the rectified linear unit, which sends negative inputs to zero. Chaining these layers gives a continuous piecewise linear function of the input. The second half counts the flat pieces produced when folds cut a space of dimension . Following a paper by Kleinberg, Strang expresses the count through binomial coefficients and a recursion formula. Session 26 of MIT 18.065 assumes matrix multiplication and function composition, nothing beyond undergraduate linear algebra.
Deep neural networks build a learning function from training data, and Strang sets up that function for binary classification of feature vectors. After a demonstration on the TensorFlow Playground, he constructs one layer from a weight matrix, a bias offset and the rectified linear unit, which sends negative inputs to zero. Chaining these layers gives a continuous piecewise linear function of the input. The second half counts the flat pieces produced when folds cut a space of dimension . Following a paper by Kleinberg, Strang expresses the count through binomial coefficients and a recursion formula. Session 26 of MIT 18.065 assumes matrix multiplication and function composition, nothing beyond undergraduate linear algebra.

This lecture
Lecture 26 of 34 of Matrix Methods in Data Analysis, Signal Processing, and Machine Learning.
Choose a node to see how it connects to this lecture. What comes before is on the leftat the top. What comes after is on the rightat the bottom. The ring is what it teaches. People and places are above. Other lectures and books are below.
Press a chosen node again to open it.
Your comment is kept while you sign up.Sign in
Kleinberg's paper