Potential-based multiobjective reinforcement learning approaches to low-impact agents for AI safety.
Saved in:
| Title: | Potential-based multiobjective reinforcement learning approaches to low-impact agents for AI safety. |
|---|---|
| Authors: | Vamplew, Peter1, p.vamplew@federation.edu.au, Foale, Cameron1, Dazeley, Richard2, Bignold, Adam1 |
| Source: | Engineering Applications of Artificial Intelligence; Apr2021, Vol. 100, pN.PAG-N.PAG, 1p |
| Database: | Applied Science & Technology Source |
| ISSN: | 09521976 |
|---|---|
| DOI: | 10.1016/j.engappai.2021.104186 |
