In This Course, We Will Study Methodologies to Guarantee The Reliability, Robustness, Equity, and Reproducibility of Advanced Machine Learning Systems, Such As Deep Neural Nets. We Will Discuss Recent Concerns About Modern Machine Learning Algorithms and Will Tackle These By Introducing Flexible Tools That Are Supported By Theoretical Guarantees. We Will Focus On Prediction Uncertainty Estimation As Well As Mitigation of Discrimination Against Minorities. We Will Also Explore Frameworks For Multiple Hypothesis Testing As Powerful Tools For Making New Scientific Descoveries And For Interpreting Complex Learning Systems. This Advanced Course Covers Influential Papers in The Field of Data Science. Learning Outcomes# At The End of The Course The Students Will Be Able To# 1. Quantify The Uncertainty in Predictions Obtained By Any Machine Learning Algorithm. 2. Analyze Whether Predictions Obtained By a Machine Learning Algorithm Are Biased Against Sub-populations. 3. Implement Data-driven Statistical Tests For a Single Hypothesis. 4. Implement Data-driven Statistical Tests For Multiple Hypotheses.

Faculty: Computer Science
|Graduate Studies

Pre-required courses

(44202 - Random Signals and 236756 - Introduction to Machine Learning) or (44202 - Random Signals and 46195 - Machine Learning) or (46195 - Machine Learning and 94412 - Probability (advanced)) or (94412 - Probability (advanced) and 236756 - Introduction to Machine Learning)

Course with no extra credit

48100 - Reliability in Modern Machine Learning