level-completabilty-x-active-learning
https://github.com/mahsabazzaz/level-completabilty-x-active-learning
Science Score: 57.0%
This score indicates how likely this project is to be science-related based on various indicators:
-
✓CITATION.cff file
Found CITATION.cff file -
✓codemeta.json file
Found codemeta.json file -
✓.zenodo.json file
Found .zenodo.json file -
✓DOI references
Found 1 DOI reference(s) in README -
○Academic publication links
-
○Academic email domains
-
○Institutional organization owner
-
○JOSS paper metadata
-
○Scientific vocabulary similarity
Low similarity (11.6%) to scientific vocabulary
Repository
Basic Info
- Host: GitHub
- Owner: MahsaBazzaz
- Language: Jupyter Notebook
- Default Branch: main
- Size: 3.76 MB
Statistics
- Stars: 2
- Watchers: 1
- Forks: 0
- Open Issues: 0
- Releases: 0
Metadata Files
README.md
Active Learning for Classifying 2D Grid-Based Level Completability
Abstract
Determining the completability of levels generated by procedural generators such as machine learning models can be challenging, as it can involve the use of solver agents that often require a significant amount of time to analyze and solve levels. Active learning is not yet widely adopted in game evaluations, although it has been used successfully in natural language processing, image and speech recognition, and computer vision, where the availability of labeled data is limited or expensive. In this paper, we propose the use of active learning for learning level completability classification. Through an active learning approach, we train deep-learning models to classify the completability of generated levels for Super Mario Bros., Kid Icarus, and a Zelda-like game. We compare active learning for querying levels to label with completability against random queries. Our results show using an active learning approach to label levels results in better classifier performance with the same amount of labeled data.
Results
Reduction in Number of labeled data needed to train the classifiers:
The results of accuracy of the classifer after each query in Tomb game:
Margin
Uncertainty
Random
Time Consumption
We compared the total time of training the 5 folds for each method in each game.
| Method | Game | Mario | Supercat | Cave | Tomb | |----------------|-------------|---------|----------|---------|---------| | Random | Time (s) | 5585.52 | 38740.73 | 2673.05 | 1516.38 | | Entropy | Time (s) | 8546.22 | 6453.80 | 1242.09 | 1173.99 | | Margin | Time (s) | 7844.88 | 41493.29 | 4238.02 | 1272.94 | | Uncertainty| Time (s) | 8271.34 | 40092.55 | 3101.18 | 951.57 |
Reference
If you found this code useful, please consider citing our work:
@INPROCEEDINGS{10333212,
author={Bazzaz, Mahsa and Cooper, Seth},
booktitle={2023 IEEE Conference on Games (CoG)},
title={Active Learning for Classifying 2D Grid-Based Level Completability},
year={2023},
volume={},
number={},
pages={1-4},
keywords={Computer vision;Uncertainty;Games;Speech recognition;Machine learning;Sampling methods;Natural language processing;video games;active learning;completability},
doi={10.1109/CoG57401.2023.10333212}}
Owner
- Name: Mahsa Bazzaz
- Login: MahsaBazzaz
- Kind: user
- Location: Boston
- Company: Northeastern University
- Website: https://www.linkedin.com/in/mahsa-bazzaz-212a41150/
- Repositories: 27
- Profile: https://github.com/MahsaBazzaz
Citation (CITATION.cff)
cff-version: 1.2.0
message: "If you use this work, please cite it as below."
authors:
- family-names: "Bazzaz"
given-names: "Mahsa"
url: "https://github.com/MahsaBazzaz/level-completabilty-x-active-learning"
preferred-citation:
type: article
authors:
- family-names: "Bazzaz"
given-names: "Mahsa"
orcid: "https://orcid.org/0009-0004-0022-9611"
- family-names: "Cooper"
given-names: "Seth"
orcid: "https://orcid.org/0000-0003-4504-0877"
title: "Active Learning for Classifying 2D Grid-Based Level Completability"
version: 2.0.4
doi: 10.1109/CoG57401.2023.10333212
date-released: 2023
GitHub Events
Total
- Push event: 1
Last Year
- Push event: 1