Generative Adversarial Symmetry Discovery
Authors: Jianke Yang, Robin Walters, Nima Dehmamy, Rose Yu
ICML 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We experiment on several tasks to demonstrate the capability of Lie GAN. Specifically, we aim to validate (1) whether Lie-GAN can discover different types of symmetries mentioned in Table 1; (2) whether the discovered symmetry, combined with existing models, can boost prediction performance. |
| Researcher Affiliation | Collaboration | 1University of California San Diego 2Northeastern University 3IBM Research. |
| Pseudocode | No | No pseudocode or algorithm blocks were found in the paper. |
| Open Source Code | Yes | Our code is available at https://github.com/Rose STL-Lab/Lie GAN. |
| Open Datasets | Yes | We use the simulated n-body trajectory dataset from Hamiltonian NN (Greydanus et al., 2019). |
| Dataset Splits | No | The train and test datasets are constructed to have different distributions so that knowledge of symmetry would be useful for generalization. The paper mentions train and test sets but does not provide specific details about validation splits or percentages for any dataset. |
| Hardware Specification | No | No specific hardware details (e.g., GPU/CPU models, memory, cloud instance types) used for experiments were found in the paper. |
| Software Dependencies | No | The paper mentions using "Hamiltonian Neural Networks" code and "Lorentz Net implementation" but does not provide specific version numbers for these or any other software dependencies (e.g., Python, PyTorch). |
| Experiment Setup | Yes | The learning rates for the discriminator and the generator are set to 0.0002 and 0.001, respectively. Lie GAN is trained adversarially for 100 epochs. |