IEEE VIS 2024 Content: Intuitive Design of Deep Learning Models through Visual Feedback

Intuitive Design of Deep Learning Models through Visual Feedback

JunYoung Choi - VIENCE Inc., Seoul, Korea, Republic of. Korea University, Seoul, Korea, Republic of

Sohee Park - VIENCE Inc., Seoul, Korea, Republic of

GaYeon Koh - Korea University, Seoul, Korea, Republic of

Youngseo Kim - VIENCE Inc., Seoul, Korea, Republic of

Won-Ki Jeong - VIENCE Inc., Seoul, Korea, Republic of. Korea University, Seoul, Korea, Republic of

Screen-reader Accessible PDF

Room: Bayshore VI

2024-10-17T18:21:00ZGMT-0600Change your timezone on the schedule page
2024-10-17T18:21:00Z
Exemplar figure, described by caption below
An example of proofreading of structural issues in a deep learning model (U-Net) using a proposed visual feedback-based no-code approach, and an example of the conventional method (code-based) corresponding to the errors present in the model.
Fast forward
Keywords

Deep learning, visual programming, explainable AI.

Abstract

In the rapidly evolving field of deep learning, traditional methodologies for designing models predominantly rely on code-based frameworks. While these approaches provide flexibility, they create a significant barrier to entry for non-experts and obscure the immediate impact of architectural decisions on model performance. In response to this challenge, recent no-code approaches have been developed with the aim of enabling easy model development through graphical interfaces. However, both traditional and no-code methodologies share a common limitation that the inability to predict model outcomes or identify issues without executing the model. To address this limitation, we introduce an intuitive visual feedback-based no-code approach to visualize and analyze deep learning models during the design phase. This approach utilizes dataflow-based visual programming with dynamic visual encoding of model architecture. A user study was conducted with deep learning developers to demonstrate the effectiveness of our approach in enhancing the model design process, improving model understanding, and facilitating a more intuitive development experience. The findings of this study suggest that real-time architectural visualization significantly contributes to more efficient model development and a deeper understanding of model behaviors.