Causality and Casual Machine Learning

This project was completed by students accepted on to the 2021 HPCC Systems Intern Program.

Curious about other projects we are offering? Take a look at our Ideas List.

Project Description

We want to evaluate the use of Causality algorithms as described in Causal Inference in Statistics by Judea Pearl et al.

Student will collaboratively identify appropriate datasets for testing, and will produce tested production code implementing the selected algorithms.

Completion of this project involves:

Survey latest techniques and research for Causal Modeling and Causal Inferencing
Develop tools for Causal Modeling including:
1. Statistical Independence and Conditional Independence Tests
2. Causal Modeling representation
3. Model Identifiability Detection
4. Interventional Calculus solver
5. Counterfactual Calculus solver
Causal Modeling User Interface
1. Identify multiple datasets for real-world analysis
2. Conduct analysis of selected datasets and compare with known ground truth where available
Publish paper(s) on the research
Test and document code for general release

By the mid term review we would expect you to have:

Mentor	Roger Dev Contact Details Backup Mentor: TBD Contact Details
Skills needed	Proficiency with probability and statistics Understanding of the basic concepts and techniques of Causality and Causal Inference Experience with Machine Learning algorithms Knowledge of ECL. Training manuals and online courses are available on the HPCC Systems website. Knowledge of distributed computing techniques Familiar with HPCC Systems Machine Learning Library Familiar with Data Pre-Processing Familiar with Github
Deliverables	Midterm TBC End of project TBC
Other resources	HPCC Systems website JIRA issue for this project: TBC Learning ECL documentation and on-line training courses. HPCC Systems Machine Learning Library Blog: HPCC Systems Machine Learning Tutorials HPCC Systems Official Github Account

We want to evaluate the use of Causality algorithms as described in “Causal Inference in Statistics” by Judea Pearl et al.

- Survey latest techniques and research for Causal Modeling and Causal Inferencing

- Develop tools for Causal Modeling including:

o Statistical Independence and Conditional Independence Tests

o Causal Modeling representation

o Model Identifiability Detection

o Interventional Calculus solver

o Counterfactual Calculus solver

o Causal Modeling User Interface

- Identify multiple datasets for real-world analysis

- Conduct analysis of selected datasets and compare with known ground truth where available

- Publish paper(s) on the research

- Test and document code for general release

The successful applicants will demonstrate:

- Proficiency with probability and statistics

- Understanding of the basic concepts and techniques of Causality and Causal Inference

- Experience with Machine Learning algorithms