Reinforcement Learning Experience Reuse with Policy Residual Representation

Authors: WenJi Zhou, Yang Yu, Yingfeng Chen, Kai Guan, Tangjie Lv, Changjie Fan, Zhi-Hua Zhou

IJCAI 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We experiment with the PRR network on a set of grid world navigation tasks, locomotion tasks, and fighting tasks in a video game. The results show that the PRR network leads to better reuse of experience and thus outperforms some state-of-the-art approaches.
Researcher Affiliation Collaboration 1National Key Laboratory for Novel Software Technology, Nanjing University, Nanjing, China 2Net Ease Fuxi AI Lab, Hangzhou, China
Pseudocode Yes Algorithm 1 Module training of Lij, Algorithm 2 Experience acquiring with PRR model, Algorithm 3 Experience reusing with PRR model
Open Source Code No The paper does not provide concrete access to source code for the methodology described. It only provides a link to a personal academic homepage, which does not explicitly host the code for this paper.
Open Datasets No The paper describes custom environments such as 'Fetch The Key tasks', 'Swimmer Gather environment', and a 'fighting video game', but does not provide concrete access information (specific links, DOIs, repositories, or formal citations) for these datasets or environments to make them publicly available or open for reproduction.
Dataset Splits No The paper does not provide specific dataset split information, such as percentages or sample counts for training, validation, or test sets, to reproduce the data partitioning.
Hardware Specification No The paper does not provide specific hardware details (e.g., GPU/CPU models, processor types, or memory amounts) used for running its experiments.
Software Dependencies No The paper only mentions 'PPO' as a reinforcement learning algorithm and 'Mujoco' as a physics engine, but does not provide specific version numbers for these or any other software dependencies.
Experiment Setup No The paper states 'All comparison algorithms use the same hyperparameters in the PPO algorithm' but does not provide specific concrete hyperparameter values (e.g., learning rate, batch size, or optimizer settings) in the main text.