Iso-Dream: Isolating and Leveraging Noncontrollable Visual Dynamics in World Models
Authors: Minting Pan, Xiangming Zhu, Yunbo Wang, Xiaokang Yang
NeurIPS 2022 | Conference PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Experiments show that Iso-Dream is effective in decoupling the mixed dynamics and remarkably outperforms existing approaches in a wide range of visual control and prediction domains. |
| Researcher Affiliation | Academia | Minting Pan Xiangming Zhu Yunbo Wang Xiaokang Yang Mo E Key Lab of Artificial Intelligence, AI Institute, Shanghai Jiao Tong University {panmt53, xmzhu76, yunbow, xkyang}@sjtu.edu.cn |
| Pseudocode | Yes | Algorithm 1: Iso-Dream (Highlight: Our modifications to behavior learning & policy deployment) |
| Open Source Code | No | Did you include the code, data, and instructions needed to reproduce the main experimental results (either in the supplemental material or as a URL)? [No] Our code is proprietary, but we will release the code once the paper is accepted. |
| Open Datasets | Yes | We quantitatively and qualitatively evaluate Iso-Dream on two reinforcement learning environments, i.e., Deep Mind Control Suite [45] and CARLA [11], and two real-world datasets for action-conditioned video prediction, i.e., BAIR robot pushing [13] and Robo Net [9]. |
| Dataset Splits | No | The paper mentions training and testing but does not explicitly provide specific percentages, counts, or predefined splits for training, validation, or test sets across the datasets used. It refers to 'random episodes' for training in Algorithm 1, but lacks detailed split information. |
| Hardware Specification | Yes | Our model is trained on Ubuntu system, 8 V100 GPUs with 32G. |
| Software Dependencies | No | The paper does not provide specific software dependencies or library versions (e.g., Python, PyTorch, CUDA versions). |
| Experiment Setup | Yes | The hyper-parameters ξ1 and ξ2 are set to 10 4 and 1, respectively. We use β1 = 1, β2 = 1 and α = 1 in Eq. (6) and τ = 5 in Eq. (7). |