Skip to content

Latest commit

 

History

History
381 lines (282 loc) · 22 KB

README.md

File metadata and controls

381 lines (282 loc) · 22 KB

GWU_DNSC 6290: Course Outline

Materials for a technical, nuts-and-bolts course about increasing transparency, fairness, robustness, and security in machine learning.

  • Lecture 1: Explainable Machine Learning Models
  • Lecture 2: Post-hoc Explanation
  • Lecture 3: Bias Testing and Remediation
  • Lecture 4: Machine Learning Security
  • Lecture 5: Machine Learning Model Debugging
  • Lecture 6: Responsible Machine Learning Best Practices
  • Lecture 7: Risk Mitigation Proposals for Language Models

Corrections or suggestions? Please file a GitHub issue.


Lecture 1: Explainable Machine Learning Models

Histogram, partial dependence, and ICE for a monotonic GBM and a credit card customer's most recent repayment status Source: Simple Explainable Boosting Machine Example

Lecture 1 Class Materials

Lecture 1 Additional Software Tools

Lecture 1 Additional Software Examples

Lecture 1 Additional Reading


Lecture 2: Post-hoc Explanation

A decision tree surrogate model forms a flow chart of a more complex monotonic GBM Source: Global and Local Explanations of a Constrained Model

Lecture 2 Class Materials

Lecture 2 Additional Software Tools

Lecture 2 Additional Software Examples

Lecture 2 Additional Reading


Lecture 3: Bias Testing and Remediation

Two hundred neural networks from a random grid search trained on the UCI Credit Card Default dataset Source: Lecture 3 Notes

Lecture 3 Class Materials

Lecture 3 Additional Software Tools

Lecture 3 Additional Software Examples

Lecture 3 Additional Reading


Lecture 4: Machine Learning Security

A cheatsheet for ML attacks Source: Responsible Machine Learning

Lecture 4 Class Materials

Lecture 4 Additional Software Tools

Lecture 4 Additional Software Examples

Lecture 4 Additional Reading


Lecture 5: Machine Learning Model Debugging

Residuals for an important feature betray a serious problem in a machine learning model. Source: Real-World Strategies for Model Debugging

Lecture 5 Class Materials

Lecture 5 Additional Software Tools

Lecture 5 Additional Software Examples

Lecture 5 Additional Reading


Lecture 6: Responsible Machine Learning Best Practices

A responsible machine learning workingflow

A Responsible Machine Learning Workflow Diagram. Source: Information, 11(3) (March 2020).

Lecture 6 Class Materials

Lecture 6 Additional Software Tools and Examples

Lecture 6 Additional Reading


Lecture 7: Risk Mitigation Proposals for Language Models

Past language model incidents.

A number of headlines and images relating to language model incidents. Source: Lecture 7 notes.

Lecture 7 Class Materials

Lecture 7 Additional Tools

Lecture 7 Additional Reading