Towards Mode Balancing of Generative Models via Diversity Weights

Loading...
Thumbnail Image

Access rights

openAccess
publishedVersion

URL

Journal Title

Journal ISSN

Volume Title

A4 Artikkeli konferenssijulkaisussa

Date

2023

Major/Subject

Mcode

Degree programme

Language

en

Pages

10

Series

Proceedings of the 14th International Conference on Computational Creativity (ICCC 2023)

Abstract

Large data-driven image models are extensively used to support creative and artistic work. Under the currently predominant distribution-fitting paradigm, a dataset is treated as ground truth to be approximated as closely as possible. Yet, many creative applications demand a diverse range of output, and creators often strive to actively diverge from a given data distribution. We argue that an adjustment of modelling objectives, from pure mode coverage towards mode balancing, is necessary to accommodate the goal of higher output diversity. We present diversity weights, a training scheme that increases a model's output diversity by balancing the modes in the training dataset. First experiments in a controlled setting demonstrate the potential of our method. We discuss connections of our approach to diversity, equity, and inclusion in generative machine learning more generally, and computational creativity specifically. An implementation of our algorithm is available at https://github.com/sebastianberns/diversity-weights.

Description

Keywords

Other note

Citation

Berns, S, Colton, S & Guckelsberger, C 2023, Towards Mode Balancing of Generative Models via Diversity Weights . in A Pease, J M Cunha, M Ackerman & D G Brown (eds), Proceedings of the 14th International Conference on Computational Creativity (ICCC 2023) ., 36, Association for Computational Creativity, International Conference on Computational Creativity, Waterloo, Ontario, Canada, 19/06/2023 . < https://computationalcreativity.net/iccc23/papers/ICCC-2023_paper_36.pdf >