Session: DTM-02-01: Artificial Intelligence in Design

Paper Number: 109087

109087 - A Task-Decomposed Ai-Aided Approach for Generative Conceptual Design 

Generative algorithm-based conceptual design has been innovatively applied as an emerging digital design paradigm for early-stage design ideation. With powerful pre-trained language models (PTLMs), designers can enter an initial prompt as a design requirement to generate descriptive natural language contents with machine's logic understanding. These machine-generated design ideas can be used as stimuli to inspire designers during design ideation. However, the scope, lack of transparency, and insufficient controllability of PTLMs may limit their effectiveness when assisting humans on a complex tasks like generative conceptual design. This generation method lacks organization, guidance, and a comprehensive understanding of design requirements. This can potentially lead to generated concepts that are mismatched or lack creativity. Inspired by the FBS model, this paper proposes a task decomposed AI-aided approach for generative conceptual design. We decompose a design requirement into three steps including functional reasoning, behavioural reasoning, and structural reasoning. Prompt templates and specification signifiers are specified for different steps to guide the PTLMs to generate reasonable results controllably. The output of each step becomes the input for the next, aiding in aggregating gains per step and embedding the selection preferences of human designers at each stage. A conceptual design experiment is conducted, and the results show that the conceptual design ideation with our method are more reasonable and creative.

Presenting Author: Boheng Wang Imperial College London

Presenting Author Biography: Boheng Wang isa PhD student at the Dyson School of Design Engineering at the Imperial College London, London, UK. His research focuses on computational creativity, design knowledge retrieval,and data-driven design. His research interests fall within AI-aided design, 3D based knowledge retrieval and combinational creativity.

Authors:

Boheng Wang Imperial College London
Haoyu Zuo Imperial College London
Zebin Cai Zhejiang University
Yuan Yin Imperial College London
Peter Childs Imperial College London
Lingyun Sun Zhejiang University
Liuqing Chen Zhejiang University