pmisiakos / causalbench-moebius

Submission to the CausalBench challenge with title: "Learning Gene Regulatory Networks under Few Root Causes Assumption.

CausalBench ICLR-23 Challenge

We provide the implementation of MÖBIUS method submitted to the ICLR 2023 CausalBench competition. Our work with title "Learning Gene Regulatory Networks under Few Root Causes Assumption" was awarded the 3rd prize in the competition and was presented in the Machine Learning for Drug Discovery workshop.

Install

pip install -r requirements.txt

Use

How to Evaluate MÖBIUS?

To run a custom graph inference function, set --model_name="custom" and --inference_function_file_path to the file path that contains your custom graph inference function (e.g. grnboost.py in this repo). You are given two starter implementations to choose from in src/, grnboost.py and dcdi.py. Your mission is to choose one of them and fine tune them to improve their performance. Hints on potential ways to improve the methods can be found directly in the code.

You should evaluate your method with the following command:

causalbench_run \
    --dataset_name weissmann_rpe1 \
    --output_directory ./output \
    --data_directory /path/to/data/storage \
    --training_regime partial_interventional \
    --partial_intervention_seed 0 \
    --fraction_partial_intervention $FRACTION \
    --model_name custom \
    --inference_function_file_path ./src/main.py \
    --subset_data 1.0 \
    --model_seed 0 \
    --do_filter

License

License

Authors

Panagiotis Misiakos, Chris Wendler and Markus Püschel

Computer Science Department, ETH Zurich.