Advanced Machine Learning (CS60073)
Instructor: Sourangshu Bhattacharya
Guest Instructor: Anirban Dasgupta, IIT Gandhinagar
TA : Abir De
Class Schedule: WED(11:00-11:55) , THURS(12:00-12:55) , FRI(08:00-08:55)
Classroom: CSE-108
Website: http://cse.iitkgp.ac.in/~sourangshu/cs60073_2016.html
Announcements:
- First Meeting: 6th Jan , Wednesday, 11:00 am
Content:
Syllabus:
Kernel Methods: Basics, Gaussian processes. Kernels on strings, trees, graphs.
Structured Learning: Multi-class SVM. Learning with structured output spaces: structSVM. Algorithms: Cutting plane algorithm.
Probabilistic Models: Probabilistic Principal Component Analysis, Latent Dirichlet allocation, Gibbs sampling, collapsed Gibbs sampling.
Temporal machine learning: Point process models, Hawkes process, learning and simulation. PAC learning.
Random Projections and sketches (Taken by Guest faculty): Introduction to concentration bounds, random projections:- basic, structured random projections, sparse projections, applications of random projections in solving regression problems, algorithms for matrix sampling and sparsifications, CUR decomposition as alternative to SVD and its applications in machine learning, streaming algorithms, sketches for estimating frequencies, heavy hitters (Count Min, Count Sketch), distinct count sketches by Flajolet/Martin, KMV.
Textbooks:
- Recent Literature.
- Machine Learning: A Probabilistic Perspective, Kevin P Murphy, MIT Press.
Prerequisites:
Syllabus of the CS60050: Machine Learning Course.
Course Material:
TBA
Assignments:
TBA