SergioIommi/DQN-2048
Deep Reinforcement Learning to Play 2048 (with Keras)
This project implements an intelligent agent that learns to play the game 2048 using deep reinforcement learning. It takes the current 2048 game board as input and determines the optimal next move (up, down, left, or right) to maximize the score and reach higher tiles. This is designed for researchers and enthusiasts exploring deep reinforcement learning techniques in game environments.
No commits in the last 6 months.
Use this if you are a machine learning researcher or student interested in seeing a practical application of deep Q-networks (DQN) for game playing and wish to experiment with different neural network architectures and training parameters.
Not ideal if you are looking for a pre-trained, ready-to-use bot to simply play 2048 at a high level without engaging in the underlying AI development, or if you prefer traditional game-solving algorithms like expectimax.
Stars
28
Forks
4
Language
Python
License
—
Category
Last pushed
Aug 12, 2018
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/SergioIommi/DQN-2048"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Compare
Higher-rated alternatives
Kulbear/endless-2048
[WIP] Aim to build an AI player for the 2048 game with a remarkable winning rate.
lesaun/2048-expectimax-ai
2048 game solved with Expectimax
FelipeMarcelino/2048-Gym
This projects aims to use reinforcement learning algorithms to play the game 2048.
qpwoeirut/2048-solver
A set of AIs for the 2048 tile-merging game. Includes an expectimax strategy that reaches 16384...
SrinidhiRaghavan/AI-2048-Puzzle
2048 puzzle using minimax algorithm