Page tree
Skip to end of metadata
Go to start of metadata

Browse Poster Wiki: 2022 Poster Award Winners, Awards Ceremony (Watch Recording from minute marker 1630), Posters by 2022 HPCC Systems InternsPosters by Academic Partners, Poster Judges, About Virtual Judging, 2022 Poster Contest Home Page, Poster Contest Previous Years

Zheyu Shen is a Masters student studying Data Science at Columbia University, USA.

Zheyu joined the HPCC Systems Intern Program to work on the Causality 2022 project led by Roger Dev. The Causality Project was new in 2021. Roger Dev has written the following blogs to support work on the project and provide information for those who would like to try out out new Causality Toolkit:

Zheyu's work involved a lot of research in an area that is still relatively new. He was tasked with designing and developing test cases to compare the causality tasks of different implementation to provide information about which are the most widely used and produce the best results. He also, assessed our own implementations focusing performance and adding more causality algorithm into our Causality Toolkit implementations. 

As well as the resources included here, read Zheyu's intern blog journal which includes a more in depth look of his work. 

Poster Abstract

Causal Inference is a method to determine the causal relationship between variables in a large system. Traditional statistic method can only explore the correlation in dataset. With the help of causal inference, we could now answer the questions such as what would happen if we intervene some variables, and what is the actual causal effect between any two variables.

The “Because” module, in the HPCC Systems Causality Framework, provides a Python module that implements all of the relevant algorithms for causal model representation, validation, causal inference, and ultimately counterfactual inference.

My work mainly focuses on three categories of causality algorithm:

  • Conditional Independence Testing
  • Causal Direction Test
  • Conditional Expectation Computation

For each of the categories above, new algorithms were tested, implemented and integrated into the “Because” module. All algorithms newly added outperform the algorithms implemented before, and have been merged into the main branch.


In this Video Recording, Zheyu provides a tour and explanation of his poster content. 

Causality Algorithm Development

Click on the poster for a larger image.

  • No labels