gzmason Goto Github PK

automatic-level-generation-for-robustreinforcement-learning

Our project focuses on the problem of generating synthetic levels of a game such that the levels can be used to learn an optimal policy for playing the game. Given a few pre-existing game levels we want to use deep generative models (like GANs) to generate new additional game levels. We will then train an RL agent on these levels to learn a generalized policy of playing the game. Our hypothesis is that training the agent with the additional levels will lead to an optimal policy that performs better than the policy learned from the few pre-existing levels. Our final objective is to learn such a generative model, train an RL agent with the generated levels and test its performance on a set of unseen game levels.

blade-demos

🍭 Blade Examples, you need it.

calendar

cc21fall1

Source files for EDAV Fall 2021 Mon / Wed Community Contribution Project

cs104-practice

deeplearning.ai-notes

吴恩达《深度学习》学习笔记(xmind)、作业代码、代码视频讲解

genshin-impact-analysis

Bookdown source files for EDAV final project

lahacks

linear_association_rplugin

mario-ai-framework

10th Anniversary Edition (Not endorsed by Nintendo)

mario-gan-levels

relation-extraction-dataset-finetuning

riotapi

risk-constrained-rl-env

self

test_git

test git and GitHub

the-building-data-genome-project

A collection of non-residential buildings for performance analysis and algorithm benchmarking