Scalable Geometric Fracture Assembly via Co-creation Space among Assemblers
Authors: Ruiyuan Zhang, Jiaxiang Liu, Zexi Li, Hao Dong, Jie Fu, Chao Wu
AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Our framework exhibits better performance on both Part Net and Breaking Bad datasets compared to existing state-of-the-art frameworks. Extensive experiments and quantitative comparisons demonstrate the effectiveness of our proposed framework... We carried out comprehensive experiments on two major geometric fracture assembly datasets: Part Net (Mo et al. 2019) and Breaking Bad (Sell an et al. 2022). Through numerous comparative experiments and ablation analyses, we compared our method with state-of-the-art works (Zhang et al. 2022; Zhan et al. 2020; Narayan, Nagar, and Raman 2022) and verified the effectiveness of our proposed framework. |
| Researcher Affiliation | Academia | 1 Zhejiang University 2 Hong Kong University of Science and Technology 3 Peking University {zhangruiyuan,zjljx,zexi.li,chao.wu}@zju.edu.cn, jiefu@ust.hk, hao.dong@pku.edu.cn |
| Pseudocode | No | No, the paper describes methods and processes using text and mathematical equations but does not include any explicitly labeled pseudocode or algorithm blocks. |
| Open Source Code | Yes | Our code is publicly available at https://github.com/Ruiyuan-Zhang/CCS. |
| Open Datasets | Yes | We evaluated our method and baselines on Part Net (Mo et al. 2019) datasets and Breaking Bad (Sell an et al. 2022, 2021). |
| Dataset Splits | Yes | We utilize the Chair dataset with default train/validation/test being split in the dataset, which includes 6,323 chairs. |
| Hardware Specification | No | No, the paper does not provide specific details about the hardware used, such as GPU or CPU models, or specific cloud/cluster configurations. |
| Software Dependencies | No | No, the paper mentions high-level architectures and models like Point Net and Transformer, but does not specify software libraries with version numbers (e.g., PyTorch 1.9, TensorFlow 2.x, or CUDA 11.1) that are needed to replicate the experiment. |
| Experiment Setup | Yes | Following (Zhang et al. 2022; Zhan et al. 2020; Narayan, Nagar, and Raman 2022), we set n = 5 in the experiment. The total loss is defined as follows: L = wc Lc + wt Lt + wr Lr + ws Ls, (9) where wc, wt, wr and ws denote the weight of different losses, which are empirically determined. The values of wt, wr and ws follow previous works (Sell an et al. 2022). Additionally, we add an ablation study for wc and C on Table 2. C is the hyperparameter of collision loss, that can be adjusted through grid search. |