Leveraging Hyperbolic Embeddings for Coarse-to-Fine Robot Design
Authors: Heng Dong, Junyu Zhang, Chongjie Zhang
ICLR 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | The extensive empirical studies on various challenging tasks sourced from Evo Gym show our approach s superior efficiency and generalization capability. |
| Researcher Affiliation | Academia | Heng Dong1 Junyu Zhang2 Chongjie Zhang3 1 IIIS, Tsinghua University 2 Huazhong University of Science and Technology 3 Washington University in St. Louis |
| Pseudocode | Yes | Algorithm 1: HERD: Hyperbolic Embeddings for Coarse-to-Fine Robot Design |
| Open Source Code | Yes | Our code is available on Git Hub*. Our research findings are fully reproducible. The source code is included in the Supplementary Material and will be made public once accepted. |
| Open Datasets | Yes | We evaluate our HERD framework on 15 tasks in the benchmark Evo Gym (Bhatia et al., 2021). |
| Dataset Splits | No | The paper states it uses Evo Gym and mentions training for '25M timesteps', but it does not specify explicit training/validation/test dataset splits (e.g., percentages or counts) or reference predefined splits for this environment. |
| Hardware Specification | Yes | Experiments are all conducted on NVIDIA GTX 2080Ti GPUs with 80 CPUs. |
| Software Dependencies | No | The paper mentions several algorithms and frameworks like PPO, Transformer, K-Means, and Horo PCA, and states that 'Our method HERD is implemented on the top of Cu Co (Wang et al., 2022)', but it does not provide specific version numbers for software dependencies such as Python, PyTorch, or other libraries. |
| Experiment Setup | Yes | Here we provide the hyperparameters needed to replicate our experiments in Table 1, and we also include our codes in the supplementary. |