Learning Physical Constraints with Neural Projections
Authors: Shuqi Yang, Xingzhe He, Bo Zhu
NeurIPS 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | 5 Experiments, We trained all the models using the Adam optimizer [48] on a single Nvidia RTX 2080Ti GPU., Figure 9: The MSE of the positions compared to the groundtruth., Table 1: Average constraint satisfications of 200 samples * 50 frames/sample predicted simulation results. |
| Researcher Affiliation | Academia | Shuqi Yang1 , Xingzhe He1, Bo Zhu1 1Dartmouth College, Computer Science Department shuqi.yang.gr@dartmouth.edu |
| Pseudocode | Yes | Algorithm 1: Iterative Neural Projection, Algorithm 2: Multi-Group Projection |
| Open Source Code | No | The paper does not provide an explicit statement about open-sourcing the code for the described methodology, nor does it provide a link to a code repository. |
| Open Datasets | No | Our training data covers a rich set of scenarios including rigid bodies, rods, articulations, collisions, contacts, and irregular domains, generated by different types of numerical simulators such as mass-spring, position-based dynamics, and rigid-body solvers. |
| Dataset Splits | No | The paper mentions 'training data' and 'test cases' but does not specify explicit training, validation, and test dataset splits with percentages or sample counts. |
| Hardware Specification | Yes | We trained all the models using the Adam optimizer [48] on a single Nvidia RTX 2080Ti GPU. |
| Software Dependencies | No | The paper states 'We implemented our learning model in Py Torch and our physics simulations in C++.' but does not provide specific version numbers for these software dependencies. |
| Experiment Setup | No | The paper states 'We refer readers to the supplementary for detailed parameter settings and animations,' indicating that specific experimental setup details such as hyperparameters are not provided in the main text. |