Explore 3D Dance Generation via Reward Model from Automatically-Ranked Demonstrations

Authors: Zilin Wang, Haolin Zhuang, Lu Li, Yinmin Zhang, Junjie Zhong, Jun Chen, Yu Yang, Boshi Tang, Zhiyong Wu

AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Empirical experiments demonstrate the effectiveness of E3D2 on the AIST++ dataset. and Empirical experiments on the AIST++ dataset (Li et al. 2021) demonstrate that the proposed E3D2 outperforms the behavior cloning (pure supervised) method across multiple metrics.
Researcher Affiliation Academia 1Shenzhen International Graduate School, Tsinghua University 2The University of Sydney 3Waseda University
Pseudocode No The paper describes the methodology steps but does not provide any formally labeled pseudocode or algorithm blocks.
Open Source Code No The paper provides a link to a 'demo page' for visualizations, but does not explicitly state that the source code for the methodology is available or provide a link to a code repository. The arXiv link is for the paper itself, not necessarily the code.
Open Datasets Yes We conduct the training and experiments on the AIST++ dataset (Li et al. 2021), which is the largest public available dataset for aligned 3D dance motions and music.
Dataset Splits No The paper states: 'In line with (Li et al. 2021; Siyao et al. 2022), we split these data into 952 sequences for training and 40 sequences for subsequent experiments.' This only specifies a training and testing split, without explicitly mentioning a separate validation set.
Hardware Specification No The paper does not provide specific details about the hardware used for running experiments, such as GPU or CPU models.
Software Dependencies No The paper mentions using 'Librosa to extract music features' but does not specify its version or any other software dependencies with version numbers.
Experiment Setup No The paper states, 'More details and hyper-parameter settings can be found in Appendices,' indicating that specific experimental setup details are not present in the main text.