10-601B Active Learning

This a lecture used in the Syllabus for Machine Learning 10-601B in Spring 2016

Slides

• slides (pdf)
• slides (pptx)

Readings

• Two Faces of Active Learning, by Sanjoy Dasgupta (especially sections 1 and 2) (link)

• Optional Advanced Readings:
  • Active Learning Literature Survey (by Burr Settles)
  • Active Learning Survey (by Balcan and Urner)

What You Should Know

• What is active learning:
  • Batch Active Learning
  • Selective Sampling and Active Learning
• Active learning could provide exponential improvements in label complexity (both theoretically and practically)!
• Common heuristics (e.g., those based on uncertainty sampling).
• Sampling bias.
• Safe Disagreement Based Active Learning Schemes.
  • Understand how they operate precisely in the realizable case (noise free scenarios).