PENERAPAN PEMBELAJARAN Q MENDALAM SEBAGAI ALGORITMA PEMBELARAJAN PENGUATAN UNTUK MELATIH AGEN PEMAIN GOMOKU
Artificial intelligence is developing very quickly. One method to develop artificial intelligence is reinforcement learning. Through reinforcement learning, an agent learns through interaction with its environment. This method can be used for many things from teaching a robot to walk to creating...
Saved in:
| Main Author: | |
|---|---|
| Format: | Final Project |
| Language: | Indonesia |
| Online Access: | https://digilib.itb.ac.id/gdl/view/85587 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Institut Teknologi Bandung |
| Language: | Indonesia |
| Summary: | Artificial intelligence is developing very quickly. One method to develop artificial
intelligence is reinforcement learning. Through reinforcement learning, an agent
learns through interaction with its environment. This method can be used for many
things from teaching a robot to walk to creating an artificial intelligence that can
play chess at a high skill level.
This research was conducted to see if one reinforcement learning algorithm, Deep Q
Learning, could be used to train an artificial intelligence to play a traditional board
game from Japan called Gomoku. Gomoku is a game that is very similar with Tictac-
toe and uses a 15x15 board. It was found that Deep Q Learning can produce
an artificial intelligence that has a very basic strategy for Gomoku for a relatively
small amount of training time. |
|---|