GANSlider: How Users Control Generative Models for Images using Multiple Sliders with and without Feedforward Information

Abstract

We investigate how multiple sliders with and without feedforward visualizations influence users' control of generative models. In an online study (N=138), we collected a dataset of people interacting with a generative adversarial network (StyleGAN2) in an image reconstruction task. We found that more control dimensions (sliders) significantly increase task difficulty and user actions. Visual feedforward partly mitigates this by enabling more goal-directed interaction. However, we found no evidence of faster or more accurate task performance. This indicates a tradeoff between feedforward detail and implied cognitive costs, such as attention. Moreover, we found that visualizations alone are not always sufficient for users to understand individual control dimensions. Our study quantifies fundamental UI design factors and resulting interaction behavior in this context, revealing opportunities for improvement in the UI design for interactive applications of generative models. We close by discussing design directions and further aspects.

arXiv comment15 pages, 10 figures, ACM CHI 2022

References in corpus (7)

• Pre-trained Models for Natural Language Processing: A Survey
• Alias-Free Generative Adversarial Networks
• Proxy Tasks and Subjective Measures Can Be Misleading in Evaluating Explainable AI Systems
• Semantic Photo Manipulation with a Generative Image Prior
• Swapping Autoencoder for Deep Image Manipulation
• A Comparative Analysis of Industry Human-AI Interaction Guidelines
• Mask-Guided Discovery of Semantic Manifolds in Generative Models

Cited by in corpus (10)

• The Metacognitive Demands and Opportunities of Generative AI
• DirectGPT: A Direct Manipulation Interface to Interact with Large Language Models
• PromptPaint: Steering Text-to-Image Generation Through Paint Medium-like Interactions
• PlantoGraphy: Incorporating Iterative Design Process into Generative Artificial Intelligence for Landscape Rendering
• Controlling AI Agent Participation in Group Conversations: A Human-Centered Approach
• GANzilla: User-Driven Direction Discovery in Generative Adversarial Networks
• DesignFromX: Empowering Consumer-Driven Design Space Exploration through Feature Composition of Referenced Products
• GANravel: User-Driven Direction Disentanglement in Generative Adversarial Networks
• BlendScape: Enabling End-User Customization of Video-Conferencing Environments through Generative AI
• StudyAlign: A Software System for Conducting Web-Based User Studies with Functional Interactive Prototypes