mansooralodhi/casualitysensitivitylab

Investigating a neural network response to input parameters using sensitivity analysis techniques.

Objective

Investigating a neural network response to input parameters using sensitivity analysis techniques.

Key Project Features

• Scalable
  • Any number of features
  • Higher order combinations
• Extensible
  • New data
  • New model
  • New techniques
• Model Framework Independent
  • Use Onnx

Core Techniques

• Boundary sensitivities
• One-Way sensitivity function
• Scenario decomposition
• Finite differences

Learning Outcome

• Understand essential mathematics for each technique.
• Build algorithm from existing formulation.
• Implement research paper.
• Analyze results.