IFT 6132 : Lectures - Winter 2018

See also the Winter 2017 version of the outline for an idea of where the class is heading.

Books on structured prediction

Part I: theory of structured prediction

Lecture 1 – 1/23 – introduction

Lecture 2 – 1/26 – structured prediction surrogate losses

Lecture 3 – 1/30 – Theory (binary classification)

Lecture 4 – 2/2 – Theory (structured prediction) - PAC-Bayes

Lecture 5 – 2/6 – Theory: generalization error bounds

Lecture 6 – 2/9 – Theory: calibrated convex surrogate losses

Lecture 7 – 2/13 – Kernels and RKHS

Lecture 8 – 2/16 – Theory: finish calibrated convex surrogate losses

Part II: convex optimization and structured prediction

Lecture 9 – 2/21 – convex optimization

Lecture 10 – 2/23 – convex optimization (II)

Lecture 11 – 2/27 – structured SVM optimization (I)

Lecture 12 – 3/2 – structured SVM optimization (II)

Lecture 13 – 3/6 – structured SVM optimization (III)

Lecture 14 – 3/9 – structured SVM optimization (IV)

Lecture 15 – 3/13 – structured SVM optimization (V)

Lecture 16 – 3/16 – FW convergence

Lecture 17 – 3/20 – FW for SVMstruct

Lecture 18 – 3/23 – CRF and variance reduced SGD

Lecture 19 – 3/27 – catalyst; RNN; learning to search; SeaRNN

Poster session – 4/27