Arshia Sobhan
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Arshia Sobhan has a total of 6 co-authors. Here are the top co-authors by number of shared publications:
- Philippe Pasquier
School of Interactive Arts + Technology, Simon Fraser University, Surrey, British Columbia, Canada
3 shared publications - Ahmed M. Abuzuraiq
School of Interactive Arts and Technology, Simon Fraser University, Surrey, British Columbia, Canada
2 shared publications - Adam Tindale
Digital Futures, OCAD University, Toronto, Ontario, Canada
1 shared publication - Alexandra Kitson
Department of Computer Science, University of Victoria, Victoria, British Columbia, Canada
1 shared publication - Kris Krüg
TheUpgrade.ai, Vancouver, Canada
1 shared publication - Lionel Ringenbach
Ucodia, Vancouver, Canada
1 shared publication
Latest Publications
- Open Access
- 12 June 2026
- Published by Association for Computing Machinery (ACM)
- p. 4202-4219
"There is Beauty in the Small": A Study of Visual Artists’ Impressions of Small Data and Model Crafting Approaches to Creative AI Work
Large-scale prompt-based AI generators have broadened access to media creation but often introduce friction in sustained creative practice. Visual artists must navigate limited personalization, narrow cultural coverage, homogenized aesthetics, and broader concerns around sustainability, bias, and data-scraping ethics. Small data and model crafting offer alternatives that center artists’ agency over data and models. We examine this approach through Autolume, a no-code, locally run tool for crafting and navigating generative models in real time. Across five focus groups with 25 visual artists in three two-day workshops, Autolume was used as a research probe to explore small-data practices. Findings show that participants viewed small data as a personal and responsible alternative to large-scale AI, but identified model training as a technical barrier, underscoring the need for improved explainability and support. We highlight design opportunities for AI art tools, including low-level controls, mapping UIs to ML processes, ecosystem integration, and small data as a pedagogical tool.
- 12 June 2026
- Published by Association for Computing Machinery (ACM)
- p. 4202-4219
Open Access"There is Beauty in the Small": A Study of Visual Artists’ Impressions of Small Data and Model Crafting Approaches to Creative AI Work
Large-scale prompt-based AI generators have broadened access to media creation but often introduce friction in sustained creative practice. Visual artists must navigate limited personalization, narrow cultural coverage, homogenized aesthetics, and broader concerns around sustainability, bias, and data-scraping ethics. Small data and model crafting offer alternatives that center artists’ agency over data and models. We examine this approach through Autolume, a no-code, locally run tool for crafting and navigating generative models in real time. Across five focus groups with 25 visual artists in three two-day workshops, Autolume was used as a research probe to explore small-data practices. Findings show that participants viewed small data as a personal and responsible alternative to large-scale AI, but identified model training as a technical barrier, underscoring the need for improved explainability and support. We highlight design opportunities for AI art tools, including low-level controls, mapping UIs to ML processes, ecosystem integration, and small data as a pedagogical tool.
- Open Access
- 3 September 2025
- Published by Association for Computing Machinery (ACM)
- p. 165-172
Autolume: A GAN-based No-Coding Small Data and Model Crafting Visual Synthesizer for Artistic Creation
- 3 September 2025
- Published by Association for Computing Machinery (ACM)
- p. 165-172
Open AccessAutolume: A GAN-based No-Coding Small Data and Model Crafting Visual Synthesizer for Artistic Creation
- Open Access
- 19 July 2024
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Computer Graphics and Interactive Techniques
- Vol. 7 (4) , 1-12
Unveiling New Artistic Dimensions in Calligraphic Arabic Script with Generative Adversarial Networks
We present an artistic exploration into calligraphic Arabic script, focusing on the nastaliq style predominant in Iran, by harnessing the affordances of Generative Adversarial Networks (GANs). Recognizing the unique challenges posed by Arabic script's cursive nature and its inadequate representation by conventional tools, our work seeks to bridge the gap between traditional calligraphy and novel technological capabilities. Two custom datasets are introduced, Nas4-60k and Nas4-60k-aug, designed to train our generative networks in producing calligraphic Arabic. Utilizing the StyleGAN2-ada architecture, our approach successfully generates stylistically coherent and high-quality calligraphic samples. These samples exhibit meaningful feature extraction and generalization of calligraphic features, extending beyond the training sets. Furthermore, our system reveals a continuous spectrum of calligraphic features through latent space interpolations, leading to the creation of dynamic, innovative artworks that blend traditional and contemporary elements of Arabic calligraphy. Drawing inspiration from the compositional form of siyah-mashq, our work culminates in multiple publicly presented artworks that exhibit a new mode of creative expression and highlight the potential of GANs in unveiling new artistic dimensions in calligraphic Arabic script.
- 19 July 2024
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Computer Graphics and Interactive Techniques
- Vol. 7 (4) , 1-12
Open AccessUnveiling New Artistic Dimensions in Calligraphic Arabic Script with Generative Adversarial Networks
We present an artistic exploration into calligraphic Arabic script, focusing on the nastaliq style predominant in Iran, by harnessing the affordances of Generative Adversarial Networks (GANs). Recognizing the unique challenges posed by Arabic script's cursive nature and its inadequate representation by conventional tools, our work seeks to bridge the gap between traditional calligraphy and novel technological capabilities. Two custom datasets are introduced, Nas4-60k and Nas4-60k-aug, designed to train our generative networks in producing calligraphic Arabic. Utilizing the StyleGAN2-ada architecture, our approach successfully generates stylistically coherent and high-quality calligraphic samples. These samples exhibit meaningful feature extraction and generalization of calligraphic features, extending beyond the training sets. Furthermore, our system reveals a continuous spectrum of calligraphic features through latent space interpolations, leading to the creation of dynamic, innovative artworks that blend traditional and contemporary elements of Arabic calligraphy. Drawing inspiration from the compositional form of siyah-mashq, our work culminates in multiple publicly presented artworks that exhibit a new mode of creative expression and highlight the potential of GANs in unveiling new artistic dimensions in calligraphic Arabic script.