Generalization Bounds for Estimating Causal Effects of Continuous Treatments

Authors: Xin Wang, Shengfei Lyu, Xingyu Wu, Tianhao Wu, Huanhuan Chen

NeurIPS 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental The effectiveness of ADMIT is empirically demonstrated in both synthetic and semi-synthetic experiments by outperforming the existing benchmarks. Our contributions in this paper are 4-fold: (4) we conduct both synthetic and semi-synthetic experiments in the continuous treatment setting, empirically demonstrating the effectiveness of ADMIT with superiority over the existing benchmarks.
Researcher Affiliation Academia Xin Wang Shengfei Lyu Xingyu Wu Tianhao Wu Huanhuan Chen University of Science and Technology of China {wz520, saintfe, xingyuwu, wutianhao8888}@mail.ustc.edu.cn hchen@ustc.edu.cn
Pseudocode Yes Algorithm 1 ADMIT: Average dose-response estimation via re-weighting schema
Open Source Code Yes The implementation of ADMIT is available at https://github.com/waxin/ADMIT
Open Datasets Yes We use one synthetic dataset and two semi-synthetic datasets, News [14, 7] and TCGA [38], to demonstrate the effectiveness of ADMIT. News [14, 7]: Fredrik Johansson, Uri Shalit, and David Sontag. Learning representations for counterfactual inference. In Proceedings of the 33rd International conference on machine learning, pages 3020 3029. PMLR, 2016. TCGA [38]: John N Weinstein, Eric A Collisson, Gordon B Mills, Kenna R Shaw, Brad A Ozenberger, Kyle Ellrott, Ilya Shmulevich, Chris Sander, and Joshua M Stuart. The cancer genome atlas pan-cancer analysis project. Nature genetics, 45(10):1113 1120, 2013.
Dataset Splits No The paper mentions 'test dataset' but does not specify training, validation, or test splits. Appendix C is mentioned for details, but the main text lacks this information.
Hardware Specification No The paper states, 'Details of these datasets and implementation can be found in Appendix C.' and in the checklist it states, 'Did you include the total amount of compute and the type of resources used (e.g., type of GPUs, internal cluster, or cloud provider)? [Yes] See Appendix C.' However, Appendix C is not provided in the given text.
Software Dependencies No The paper states, 'Details of these datasets and implementation can be found in Appendix C.' and in the checklist it states, 'Did you specify all the training details (e.g., data splits, hyperparameters, how they were chosen)? [Yes] See Appendix C.' However, Appendix C is not provided in the given text, and no software dependencies with versions are listed in the main text.
Experiment Setup No The paper states, 'Details of these datasets and implementation can be found in Appendix C.' and in the checklist it states, 'Did you specify all the training details (e.g., data splits, hyperparameters, how they were chosen)? [Yes] See Appendix C.' However, Appendix C is not provided in the given text.