BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Logistic and Softmax Regression\, and their Relation to the Neural
  Network World - Adrian Scoica (University of Cambridge)
DTSTART:20140519T130000Z
DTEND:20140519T140000Z
UID:TALK52337@talks.cam.ac.uk
CONTACT:Advait Sarkar
DESCRIPTION:Logistic Regression\, along with its generalized counterpart S
 oftmax\nRegression\, is one of the most popular and best-performing genera
 lized\nlinear classification algorithms currently used in Machine Learning
 .\nIn this lecture\, I will explain some of the intuitions behind using\na
 nd training these classifiers\, and I will show how they are related\nto N
 eural Networks.\n\nUsing cleverly assembled examples from the harsh and un
 forgiving world\nof dating\, I will reveal the probabilistic concepts that
  lie behind\nthe logistic function. I will demonstrate how to train a bina
 ry\nlogistic regression classifier using gradient descent\, and I will sho
 w\nhow those intuitions generalize naturally to the multi-class problem.\n
 Last but not least\, we will see how these classifiers can be thought\nof 
 as very simple Artificial Neural Networks\, and thus can be used as\nlayer
  components in more complicated Neural Network architectures.\n\nThe slide
 s can be viewed "here":https://docs.google.com/presentation/d/1xUKFeOOXC9P
 iuZKjiBKXqoXp2h2Be3kqJzJrrtibirQ/edit?usp=sharing.
LOCATION:LT2\, Computer Laboratory\, William Gates Building
END:VEVENT
END:VCALENDAR