Personalized Reinforcement Learning with a Budget of Policies

Authors: Dmitry Ivanov, Omer Ben-Porat

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

Reproducibility Variable Result LLM Response
Research Type Experimental Our empirical investigations, conducted across a variety of simulated environments, showcase the algorithms ability to facilitate meaningful personalization even under constrained policy budgets.
Researcher Affiliation Academia Dmitry Ivanov, Omer Ben-Porat Technion, Israel divanov@campus.technion.ac.il, omerbp@technion.ac.il
Pseudocode No The paper describes its algorithms in prose (e.g., 'Our EM-like algorithm is inspired by...', 'Our second algorithm is based on an observation...') but does not provide formal pseudocode blocks or algorithm figures.
Open Source Code Yes Code: https://github.com/dimonenka/RL_policy_budget
Open Datasets Yes We use the Resource Gathering environment adapted from (Barrett and Narayanan 2008; Alegre et al. 2022), where a policy directs a character in a 5x5 grid world to collect resources. To rigorously test our algorithms in more complex scenarios, we employ Mu Jo Co environments (Todorov, Erez, and Tassa 2012; Tassa et al. 2018; Tunyasuvunakool et al. 2020), including Half Cheetah, Ant, Hopper, and Walker2d.
Dataset Splits No The paper does not explicitly specify a validation dataset or a distinct validation split for its experiments. It mentions training and evaluating performance but not a separate validation phase with specified data splits.
Hardware Specification No The paper does not mention any specific hardware components (e.g., GPU models, CPU models, memory specifications) used for conducting the experiments.
Software Dependencies No The paper does not specify version numbers for any key software dependencies (e.g., Python, PyTorch/TensorFlow, specific libraries or frameworks) used in the implementation of their algorithms or experiments.
Experiment Setup No The main text of the paper states: 'To ensure reproducibility and transparency, hyperparameters and technical details are provided in the Appendix.' However, the Appendix is not provided in the scope of this analysis, so the specific experimental setup details are not available in the main text.