Human-centred mechanism design with Democratic AI

Building artificial intelligence (AI) that aligns with human values is an unsolved problem. Here we developed a human-in-the-loop research pipeline called Democratic AI, in which reinforcement learning is used to design a social mechanism that humans prefer by majority. A large group of humans playe...

Descripción completa

Detalles Bibliográficos
Autores principales: Koster, Raphael, Balaguer, Jan, Tacchetti, Andrea, Weinstein, Ari, Zhu, Tina, Hauser, Oliver, Williams, Duncan, Campbell-Gillingham, Lucy, Thacker, Phoebe, Botvinick, Matthew, Summerfield, Christopher
Formato: Online Artículo Texto
Lenguaje:English
Publicado: Nature Publishing Group UK 2022
Materias:
Article
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9584820/
https://www.ncbi.nlm.nih.gov/pubmed/35789321
http://dx.doi.org/10.1038/s41562-022-01383-x

Internet

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9584820/
https://www.ncbi.nlm.nih.gov/pubmed/35789321
http://dx.doi.org/10.1038/s41562-022-01383-x

Ejemplares similares