Projects in Machine Learning: Selected Topics

Overview

In this research-oriented project course, students propose and carry out projects on advanced topics in machine learning. There will be several introductory tutorial lectures providing background and highlighting possible project ideas. Student groups will be mentored by the instructors, and present their progress to the class. Topics of interest include, among others:

Topics covered

• Online decision making: Online learning, multi-armed bandits, reinforcement learning, exploration/exploitation tradeoffs, learning under partial observability.
• Active learning: Bayesian experimental design, adaptive sampling, optimized information gathering
• Combinatorial approaches: structured output prediction, discrete optimization in machine learning, sparsity, submodular functions

Details

• Lecturers:
• Prof. Dr. Andreas Krause (CAB G68)
  
• Dr. Gábor Bartók (CAB F65.1)
  
• Dr. Amin Karbasi (CAB F63.2)
  
• VVZ Information: See here.
• Resources
  
• Lecture:
  • Tuesday 15-16 in CAB G59
  • Thursday 15-17 in CAB G51

Slides

• 27.2, 4.3.: Submodularity tutorial (Andreas Krause) [pdf]. More references available here.
  
• 20.2., 25.2.: Online learning tutorial (Gabor Bartok) [pdf]
  
• 18.2.: Introduction [pdf]