Collaborative Causal Discovery with Atomic Interventions
Authors: Raghavendra Addanki, Shiva Kasiviswanathan
NeurIPS 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | In Section 5, we show experiments on data generated from both real and synthetic networks with added latents and demonstrate the efficacy of our algorithms for learning the underlying clustering and the MAGs. |
| Researcher Affiliation | Collaboration | Raghavendra Addanki University of Massachusetts, Amherst raddanki@cs.umass.edu Shiva Prasad Kasiviswanathan Amazon kasivisw@gmail.com |
| Pseudocode | Yes | Algorithm 1 IDENTIFY-OUTNBR (Ui, u) |
| Open Source Code | No | The paper does not provide any explicit statements about releasing source code, nor does it include a link to a code repository. |
| Open Datasets | Yes | We consider the following real-world Bayesian networks from the Bayesian Network Repository which cover a wide variety of domains: Asia (Lung cancer) (8 nodes, 8 edges), Earthquake (5 nodes, 4 edges), Sachs (Protein networks) (11 nodes, 17 edges), and Survey (6 nodes, 6 edges). For the synthetic data, we use Erdös-Rényi random graphs (10 nodes). |
| Dataset Splits | No | The paper mentions generating data from real and synthetic networks for experiments and evaluating clustering performance, but it does not specify explicit train/validation/test dataset splits, percentages, or absolute sample counts. |
| Hardware Specification | No | The paper does not provide any specific details regarding the hardware (e.g., GPU/CPU models, memory) used for running the experiments. |
| Software Dependencies | No | The paper does not specify any software dependencies with version numbers (e.g., programming languages, libraries, or specific software packages) used for the experiments. |
| Experiment Setup | Yes | We set number of entities M = 40, number of clusters k = 2, α = 0.60, β = 0.20, and dominant MAG parameter γ = 0.90 for both the clusters. For the synthetic data generated using Erdös-Rényi model, we use n = 10, probability of edge 0.3. |