Transfer Learning on Heterogeneous Feature Spaces for Treatment Effects Estimation

Ioana Bica, Mihaela van der Schaar

Neural Information Processing Systems (NeurIPS) 2022

HTCE-learners

PyTorch implementation for the HTCE-learners.

Note that the code in contrib/ for CATENets [1, 2] is from: https://github.com/AliciaCurth/CATENets and the code for RadialGAN [3] is adapted from: https://github.com/vanderschaarlab/mlforhealthlabpub/tree/main/alg/RadialGAN

Install requirements from requirements.txt. Python version 3.6 or 3.8 recommended. See comments inside requirements.txt for additional installation notes.

To reproduce the experiments in the paper (for the Twins dataset), use the following commands:

• Benchmark comparison (Table 1)

python run_experiments.py --experiment_name="baseline_experiment"

• Varying the information sharing between domains (Figure 6 - top)

python run_experiments.py --experiment_name="po_sharing_across_domains"

• Varying the target dataset size (Figure 6 - bottom)

python run_experiments.py --experiment_name="target_size"

• Effect of selection bias (Figure 7)

python run_experiments.py --experiment_name="selection_bias"

The results are saved in results/. To plot the results from the paper, use the Jupyter notebook in results/results_figs/analyze_results.ipynb.

References

[1] Curth, Alicia, and Mihaela van der Schaar. "Nonparametric estimation of heterogeneous treatment effects: From theory to learning algorithms". International Conference on Artificial Intelligence and Statistics. PMLR, 2021.

[2] Curth, Alicia, and Mihaela van der Schaar. "On inductive biases for heterogeneous treatment effect estimation". Advances in Neural Information Processing Systems, 2021.

[3] Yoon, Jinsung, James Jordon, and Mihaela Schaar. "RadialGAN: Leveraging multiple datasets to improve target-specific predictive models using Generative Adversarial Networks". International Conference on Machine Learning. PMLR, 2018.