Estimating Possible Causal Effects with Latent Variables via Adjustment
Authors: Tian-Zuo Wang, Tian Qin, Zhi-Hua Zhou
ICML 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiments validate the effectiveness and tremendous efficiency improvement of the proposed method. [...] We generate random DAGs with vertex number d = 8, 10, 12, 14, 16 and each edge between two vertices occurs with probability ρ = 0.2, 0.3, 0.4, 0.5, which is called graph density. [...] We show the average number of the output set of causal effects and running time in 100 graphs under each parameter in Fig. 3. |
| Researcher Affiliation | Academia | 1National Key Laboratory for Novel Software Technology, Nanjing University, Nanjing, 210023, China. Correspondence to: Zhi-Hua Zhou <zhouzh@lamda.nju.edu.cn>. |
| Pseudocode | Yes | Algorithm 1: PAGcauses |
| Open Source Code | No | The paper does not contain any explicit statements about releasing source code or provide links to a code repository. |
| Open Datasets | No | We generate random DAGs with vertex number d = 8, 10, 12, 14, 16 and each edge between two vertices occurs with probability ρ = 0.2, 0.3, 0.4, 0.5, which is called graph density. [...] We obtain PAG directly with the true covariance matrix of the observable variables. |
| Dataset Splits | No | For each set of parameters, we generate 100 causal graphs and obtain the output set in the time limit of 3000 seconds for each graph. The paper describes generating multiple graphs for simulations rather than splitting a single dataset into train/validation/test sets. |
| Hardware Specification | No | The paper does not specify any hardware details (e.g., CPU, GPU models, memory, or specific computing environments) used for running the experiments. |
| Software Dependencies | No | The paper does not mention any specific software dependencies or their version numbers. |
| Experiment Setup | Yes | We generate random DAGs with vertex number d = 8, 10, 12, 14, 16 and each edge between two vertices occurs with probability ρ = 0.2, 0.3, 0.4, 0.5, which is called graph density. The DAG is parameterized as a linear Gaussian structural equation model. The weight of each edge is drawn from Uniform([1, 2]). [...] The maximum running time for each simulation is 3000s. |