1 of 8

Jing Nathan Yan, Ziwei Gu, Hubert Lin, Jeffrey M. Rzeszotarski

SILVA: Interactively Assessing Machine Learning Fairness Using Causality

2 of 8

SILVA

3 of 8

Existing

Machine

Learning Fairness Assessment Tool

4 of 8

Key Contributions

1. A tool for machine learning fairness assessment using causality and free-form exploration

2. Novel causal graph visualization and user interaction design for what-if queries

3. Extensive user studies to demonstrate the effectiveness of Silva over an existing system (AIF360)

5 of 8

Key Contributions

1. A tool for machine learning fairness assessment using causality and free-form exploration

2. Novel causal graph visualization and user interaction design for what-if queries

3. Extensive user studies to demonstrate the effectiveness of Silva over an existing system (AIF360)

6 of 8

Key Contributions

1. A tool for machine learning fairness assessment using causality and free-form exploration

2. Novel causal graph visualization and user interaction design for what-if queries

3. Extensive user studies to demonstrate the effectiveness of Silva over an existing system (AIF360)

7 of 8

Key Contributions

1. A tool for machine learning fairness assessment using causality and free-form exploration

2. Novel causal graph visualization and user interaction design for what-if queries

3. Extensive user studies to demonstrate the effectiveness of Silva over an existing system (AIF360)

Better Bias Discovery Rate

Better User Experience

8 of 8

Key Contributions

1. A tool for machine learning fairness assessment using causality and free-form exploration

2. Novel causal graph visualization and user interaction design for what-if queries

3. Extensive user studies to demonstrate the effectiveness of Silva over comparable systems