Improving Generative Imagination in Object-Centric World Models
Authors: Zhixuan Lin, Yi-Fu Wu, Skand Peri, Bofeng Fu, Jindong Jiang, Sungjin Ahn
ICML 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Our thorough investigation on the temporal generation ability in comparison to the previous models demonstrates that G-SWM achieves the versatility with the best or comparable performance for all experiment settings including a few complex settings that have not been tested before. We evaluate G-SWM on several datasets designed to illus trate generation quality with respect to the different abilities outlined in Table 1. |
| Researcher Affiliation | Academia | 1Rutgers Uni versity 2Zhejiang University 3Tianjin University 4Rutgers Cen ter for Cognitive Science. |
| Pseudocode | No | The paper does not contain structured pseudocode or algorithm blocks. |
| Open Source Code | No | The paper provides a URL (https://sites.google.com/view/gswm) which is a project page, not a direct link to a code repository, and does not explicitly state that source code for the methodology is released at this link or in supplementary materials. |
| Open Datasets | Yes | We use the CLEVR (Johnson et al., 2016) dataset as a starting point to create a dynamic 3D environment. |
| Dataset Splits | Yes | Each episode contains 100 timesteps and we train the models on random sequences of length 20. For this experi ment, we train on sequences of length 10. |
| Hardware Specification | No | The paper does not provide specific hardware details (e.g., GPU/CPU models, memory) used for running its experiments. |
| Software Dependencies | No | The paper does not provide specific ancillary software details with version numbers (e.g., library or solver names with version numbers like Python 3.8, PyTorch 1.9). |
| Experiment Setup | Yes | Related hyperparameters are provided in Appendix. Each episode contains 100 timesteps and we train the models on random sequences of length 20. For this experi ment, we train on sequences of length 10 and at test time provide 5 ground truth steps to generate the following steps. |