Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in [1].
Efficiently Adapt to New Dynamic via Meta-Model
Authors: Kaixin Huang, Chen Zhao, Chun Yuan
JAIR 2024 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Through extensive experimentation encompassing diverse simulated robotics and control tasks, we validate the efficacy of our approach and demonstrate its superior generalization ability compared to existing schemes, and explore multiple strategies for obtaining policies with personalized models. Our method achieves a model with reduced prediction error, outperforming previous methods in policy performance, and facilitating efficient adaptation when compared to prior dynamic model generalization methods and OMRL algorithms. |
| Researcher Affiliation | Academia | Kaixin Huang EMAIL Tsinghua Shenzhen International Graduate School, Shenzhen, Guangdong, China Chen Zhao EMAIL Tsinghua Shenzhen International Graduate School, Shenzhen, Guangdong, China Chun Yuan EMAIL Tsinghua Shenzhen International Graduate School, Shenzhen, Guangdong, China |
| Pseudocode | Yes | Appendix A. PSEUDO-CODE Algorithm 1 Training Phase(transition function) Algorithm 2 Adaptation Phase(transition function) |
| Open Source Code | No | The paper does not contain any explicit statements about releasing their own source code, nor does it provide a link to a code repository. It only mentions using third-party tools and environments. |
| Open Datasets | Yes | We evaluate PLDMM on various benchmark environments from Open AI Gym (Brockman et al., 2016) and Mu Jo Co physics engine (Todorov et al., 2012). |
| Dataset Splits | Yes | Using the action selection procedure described in (ii), 100 trajectories were collected in each training environment for the purpose of training the models. Additionally, 5 trajectories were collected in each testing task for the adaptation phase. |
| Hardware Specification | No | The paper does not provide specific hardware details such as CPU or GPU models, or cloud computing instance types, used for running the experiments. |
| Software Dependencies | No | The paper mentions using a 'Transformer Pytorch implementation' and environments like 'Open AI gym' and 'Mu Jo Co', but it does not specify exact version numbers for these or other software libraries/dependencies. |
| Experiment Setup | Yes | In this section, we list the hyperparameters in the training phases that we used to produce the experimental results when compared with model generalization methods Table 7 or OMRL methods Table 8. Table 7: hyperparameters in the training phases when compared with model generalization methods. Table 8: hyperparameters in the training phases when compared with OMRL methods. |