Evolving Mario levels in the latent space of a deep convolutional generative adversarial network
- Submitting institution
-
Queen Mary University of London
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 505
- Type
- E - Conference contribution
- DOI
-
10.1145/3205455.3205517
- Title of conference / published proceedings
- GECCO 2018 - Proceedings of the 2018 Genetic and Evolutionary Computation Conference
- First page
- 221
- Volume
- -
- Issue
- -
- ISSN
- -
- Open access status
- -
- Month of publication
- July
- Year of publication
- 2018
- URL
-
-
- Supplementary information
-
-
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- No
- Number of additional authors
-
5
- Research group(s)
-
-
- Citation count
- 11
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- Novel approach demonstrated how the latent space induced by GANs is efficiently exploitable by evolutionary search to generate novel content to meet specified objectives. An international collaboration between researchers from the UK, Denmark, Germany, China and the USA from a working group at Dagstuhl seminar (17471-2017) on AI and Games. Won a best paper award at GECCO 2018 in Kyoto (+ follow-up papers GECCO 2019, 2020). Featured in Science Magazine (https://www.sciencemag.org/news/2018/05/bored-your-video-game-artificial-intelligence-could-create-new-levels-fly ). Volz (a co-author) won the 2018 IEEE CIG Conference Short Video prize (on topic), joined Game-AI startup modl.ai (Copenhagen). Software open source (https://github.com/schrum2/GameGAN ).
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -
