GP-for-interpretable-RL
所属分类：人工智能/神经网络/深度学习
开发工具：TeX
文件大小：1370KB
下载次数：0
上传日期：2020-10-31 23:28:36
上 传 者：
sh-1993
说明： 遗传编程算法在流行的RL基准上训练，使解更容易解释。
(Genetic Programming algorithm trained on popular RL benchmarks to make solutions more interpretable.)
文件列表:
LICENSE (1078, 2020-11-01)
animations (0, 2020-11-01)
animations\cartpole_random.gif (117589, 2020-11-01)
animations\cartpole_solved.gif (331990, 2020-11-01)
animations\mountaincar_random.gif (863655, 2020-11-01)
animations\mountaincar_solved.gif (596184, 2020-11-01)
animations\pendulum_solved.gif (713964, 2020-11-01)
environment.yml (130, 2020-11-01)
report (0, 2020-11-01)
report\background.tex (16920, 2020-11-01)
report\cartpole_mountaincar.tex (17152, 2020-11-01)
report\conclusion.tex (2532, 2020-11-01)
report\images (0, 2020-11-01)
report\images\cartpole.png (4731, 2020-11-01)
report\images\complex_iflte_program.png (4834, 2020-11-01)
report\images\mountain-car.png (10010, 2020-11-01)
report\images\mountaincarcont.png (14013, 2020-11-01)
report\images\neural_network.png (104370, 2020-11-01)
report\images\pend_simple_gp_agent.png (22032, 2020-11-01)
report\images\pendulum.png (14303, 2020-11-01)
report\images\pendulum_deep_simple_GP.png (27618, 2020-11-01)
report\images\pendulum_quadrants.png (26198, 2020-11-01)
report\images\simple_iflte_program.png (4348, 2020-11-01)
report\introduction.tex (6335, 2020-11-01)
report\main.tex (2331, 2020-11-01)
report\methods.tex (9210, 2020-11-01)
report\pendulum.tex (27079, 2020-11-01)
report\professional_considerations.tex (1463, 2020-11-01)
report\references.bib (5243, 2020-11-01)
report\report.pdf (495149, 2020-11-01)
src (0, 2020-11-01)
src\cartpole.py (3673, 2020-11-01)
src\cartpole_agent_old.py (5029, 2020-11-01)
src\cartpole_binary_agent.py (3241, 2020-11-01)
src\cartpole_binary_agent_exp.py (383, 2020-11-01)
src\cartpole_info.py (766, 2020-11-01)
src\gp_gym_info.py (301, 2020-11-01)
... ...
# Overview
This project is part of my undergraduate dissertation BSc Computer Science at the University of Sussex.
You can read the dissertation [here](https://alexgeorgousis.github.io/GP-for-interpretable-RL/report/report.pdf).
# Project Aim
The aim of the project is to use Genetic Programming to produce interpretable solutions to popular Reinforcement Learning tasks.
近期下载者：
相关文件：
收藏者：