Generating Behaviorally Diverse Policies with Latent Diffusion Models

doi:doi:10.57702/drl7n1nd

Generating Behaviorally Diverse Policies with Latent Diffusion Models

Quality Diversity (QD) is an emerging field in which collections of high performing, behaviorally diverse solutions are trained. The foundational method, Map Elites, maintains an archive of solutions where each cell in the archive corresponds to a solution with a score given by the task objective f, and behavior specified by measure functions m1...mk, which map to a low dimensional behavior space.

BibTex:

@dataset{Shashank_Hegde_and_Sumeet_Batra_and_KR_Zentner_and_Gaurav_S_Sukhatme_2024,
    abstract = {Quality Diversity (QD) is an emerging field in which collections of high performing, behaviorally diverse solutions are trained. The foundational method, Map Elites, maintains an archive of solutions where each cell in the archive corresponds to a solution with a score given by the task objective f, and behavior specified by measure functions m1...mk, which map to a low dimensional behavior space.},
    author = {Shashank Hegde and Sumeet Batra and K.R. Zentner and Gaurav S. Sukhatme},
    doi = {10.57702/drl7n1nd},
    institution = {No Organization},
    keyword = {'archive distillation', 'latent diffusion models', 'quality diversity reinforcement learning'},
    month = {dec},
    publisher = {TIB},
    title = {Generating Behaviorally Diverse Policies with Latent Diffusion Models},
    url = {https://service.tib.eu/ldmservice/dataset/generating-behaviorally-diverse-policies-with-latent-diffusion-models},
    year = {2024}
}