Demystifying Linear MDPs and Novel Dynamics Aggregation Framework

Authors: Joongkyu Lee, Min-hwan Oh

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

Reproducibility Variable Result LLM Response
Research Type Experimental In numerical experiments, our proposed method consistently outperforms existing algorithms by significant margins. Our main contributions can be summarized as follows:
Researcher Affiliation Academia Joongkyu Lee Graduate School of Data Science Seoul National University jklee0717@snu.ac.kr Min-hwan Oh Graduate School of Data Science Seoul National University minoh@snu.ac.kr
Pseudocode Yes Algorithm 1 Upper Confidence Hierarchical RL with Transition-Targeted Regression (UC-HRL)
Open Source Code No The paper does not provide a concrete statement or link for the availability of its source code.
Open Datasets No The paper describes the "Block-River Swim" environment in Appendix H, which is a variant of River Swim (Strehl & Littman, 2008). This describes the simulation environment but does not provide access information (link, citation with authors/year for a public dataset, or explicit statement of public availability) for a static dataset used in experiments.
Dataset Splits No The paper mentions running "Episodic returns over 10 independent runs" but does not specify details of train/validation/test splits, percentages, or predefined splits for any dataset used.
Hardware Specification No The paper does not provide specific details about the hardware used for running experiments (e.g., GPU/CPU models, memory).
Software Dependencies No The paper does not provide specific software dependencies with version numbers.
Experiment Setup No The paper states, "For a fair comparison, we sweep over the hyper-parameters for each algorithm over certain ranges." However, it does not list specific hyperparameter values, training configurations, or system-level settings in the main text.