Deep Learning: Theory and Practice Matrix Calculus 31-1-2019 - - PowerPoint PPT Presentation

▶

Oct 25, 2023 153 likes •265 views

Deep Learning: Theory and Practice Matrix Calculus 31-1-2019 Linear and Logistic Regression Models deeplearning.cce2019@gmail.com Matrix Derivatives Linear Models for Classification Optimize a modified cost function Bishop - PRML book

SLIDE 1

Deep Learning: Theory and Practice

31-1-2019

Matrix Calculus Linear and Logistic Regression Models

deeplearning.cce2019@gmail.com

SLIDE 2

Matrix Derivatives

SLIDE 3

Linear Models for Classification

Bishop - PRML book (Chap 3)

❖ Optimize a modified cost function

SLIDE 4

Least Squares for Classification

Bishop - PRML book (Chap 3)

❖ K-class classification problem ❖ With 1-of-K hot encoding, and

least squares regression

SLIDE 5

Gradient Descent For Function Minimization

SLIDE 6

Non-linear Optimization

Typical Error Surface as a function of parameters (weights) Highly Non-linear

SLIDE 7

Approximate Minimization

SLIDE 8

Approximate Minimization

Error surface close to a local optima Move to local optima

Method of Steepest Descent

SLIDE 9

Logistic Regression

Bishop - PRML book (Chap 3)

❖ 2- class logistic regression ❖ K-class logistic regression ❖ Maximum likelihood solution ❖ Maximum likelihood solution

SLIDE 10

Least Squares versus Logistic Regression

Bishop - PRML book (Chap 4)

SLIDE 11

Least Squares versus Logistic Regression

Bishop - PRML book (Chap 4)