Generative Pre-Trained Transformer for Design Concept Generation: An Exploration

Q. Zhu; J. Luo

doi:10.1017/pds.2022.185

Generative Pre-Trained Transformer for Design Concept Generation: An Exploration

Q. Zhu^*, J. Luo

^*Corresponding author for this work

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

90 Citations (Scopus)

40 Downloads (CityUHK Scholars)

Abstract

Novel concepts are essential for design innovation and can be generated with the aid of data stimuli and computers. However, current generative design algorithms focus on diagrammatic or spatial concepts that are either too abstract to understand or too detailed for early phase design exploration. This paper explores the uses of generative pre-trained transformers (GPT) for natural language design concept generation. Our experiments involve the use of GPT-2 and GPT-3 for different creative reasonings in design tasks. Both show reasonably good performance for verbal design concept generation. © The Author(s), 2022.

Original language	English
Title of host publication	INTERNATIONAL DESIGN CONFERENCE - DESIGN 2022
Editors	Mario Štorga, Stanko Škec, Tomislav Martinec, Dorian Marjanović
Publisher	Cambridge University Press
Pages	1825-1834
DOIs	https://doi.org/10.1017/pds.2022.185
Publication status	Published - May 2022
Externally published	Yes
Event	17th International Design Conference (DESIGN 2022) - Virtual, Croatia Duration: 23 May 2022 → 26 May 2022 https://www.designconference.org/past-events

Publication series

Name	Proceedings of the Design Society
Volume	2
ISSN (Print)	2732-527X

Conference

Conference	17th International Design Conference (DESIGN 2022)
Place	Croatia
Period	23/05/22 → 26/05/22
Internet address	https://www.designconference.org/past-events

Research Keywords

early design phase
generative design
generative pre-trained transformer
idea generation
natural language generation

Publisher's Copyright Statement

This full text is made available under CC-BY-NC-ND 4.0. https://creativecommons.org/licenses/by-nc-nd/4.0/

Access Output

10.1017/pds.2022.185

192712984.pdfFinal Published version, 613 KBLicence: CC BY-NC-ND

Other Access Options

Check CityUHK Library

Permanent Link

Right click to copy link

Cite this

@inproceedings{6980343fb5a149afa52f2ea203ebddab,

title = "Generative Pre-Trained Transformer for Design Concept Generation: An Exploration",

abstract = "Novel concepts are essential for design innovation and can be generated with the aid of data stimuli and computers. However, current generative design algorithms focus on diagrammatic or spatial concepts that are either too abstract to understand or too detailed for early phase design exploration. This paper explores the uses of generative pre-trained transformers (GPT) for natural language design concept generation. Our experiments involve the use of GPT-2 and GPT-3 for different creative reasonings in design tasks. Both show reasonably good performance for verbal design concept generation. {\textcopyright} The Author(s), 2022.",

keywords = "early design phase, generative design, generative pre-trained transformer, idea generation, natural language generation",

author = "Q. Zhu and J. Luo",

year = "2022",

month = may,

doi = "10.1017/pds.2022.185",

language = "English",

series = "Proceedings of the Design Society",

publisher = "Cambridge University Press",

pages = "1825--1834",

editor = "Mario {\v S}torga and Stanko {\v S}kec and Tomislav Martinec and Dorian Marjanovi{\'c}",

booktitle = "INTERNATIONAL DESIGN CONFERENCE - DESIGN 2022",

address = "United Kingdom",

note = "17th International Design Conference (DESIGN 2022) ; Conference date: 23-05-2022 Through 26-05-2022",

url = "https://www.designconference.org/past-events",

}

Zhu, Q & Luo, J 2022, Generative Pre-Trained Transformer for Design Concept Generation: An Exploration. in M Štorga, S Škec, T Martinec & D Marjanović (eds), INTERNATIONAL DESIGN CONFERENCE - DESIGN 2022. Proceedings of the Design Society, vol. 2, Cambridge University Press, pp. 1825-1834, 17th International Design Conference (DESIGN 2022), Croatia, 23/05/22. https://doi.org/10.1017/pds.2022.185

Generative Pre-Trained Transformer for Design Concept Generation: An Exploration. / Zhu, Q.; Luo, J.
INTERNATIONAL DESIGN CONFERENCE - DESIGN 2022. ed. / Mario Štorga; Stanko Škec; Tomislav Martinec; Dorian Marjanović. Cambridge University Press, 2022. p. 1825-1834 (Proceedings of the Design Society; Vol. 2).

Research output: Chapters, Conference Papers, Creative and Literary Works › RGC 32 - Refereed conference paper (with host publication) › peer-review

TY - GEN

T1 - Generative Pre-Trained Transformer for Design Concept Generation

T2 - 17th International Design Conference (DESIGN 2022)

AU - Zhu, Q.

AU - Luo, J.

PY - 2022/5

Y1 - 2022/5

N2 - Novel concepts are essential for design innovation and can be generated with the aid of data stimuli and computers. However, current generative design algorithms focus on diagrammatic or spatial concepts that are either too abstract to understand or too detailed for early phase design exploration. This paper explores the uses of generative pre-trained transformers (GPT) for natural language design concept generation. Our experiments involve the use of GPT-2 and GPT-3 for different creative reasonings in design tasks. Both show reasonably good performance for verbal design concept generation. © The Author(s), 2022.

AB - Novel concepts are essential for design innovation and can be generated with the aid of data stimuli and computers. However, current generative design algorithms focus on diagrammatic or spatial concepts that are either too abstract to understand or too detailed for early phase design exploration. This paper explores the uses of generative pre-trained transformers (GPT) for natural language design concept generation. Our experiments involve the use of GPT-2 and GPT-3 for different creative reasonings in design tasks. Both show reasonably good performance for verbal design concept generation. © The Author(s), 2022.

KW - early design phase

KW - generative design

KW - generative pre-trained transformer

KW - idea generation

KW - natural language generation

UR - http://www.scopus.com/inward/record.url?scp=85131376169&partnerID=8YFLogxK

UR - https://www.scopus.com/record/pubmetrics.uri?eid=2-s2.0-85131376169&origin=recordpage

UR - https://app-overton-io.ezproxy.cityu.edu.hk/articles.php?query=10.1017/pds.2022.185

U2 - 10.1017/pds.2022.185

DO - 10.1017/pds.2022.185

M3 - RGC 32 - Refereed conference paper (with host publication)

T3 - Proceedings of the Design Society

SP - 1825

EP - 1834

BT - INTERNATIONAL DESIGN CONFERENCE - DESIGN 2022

A2 - Štorga, Mario

A2 - Škec, Stanko

A2 - Martinec, Tomislav

A2 - Marjanović, Dorian

PB - Cambridge University Press

Y2 - 23 May 2022 through 26 May 2022

ER -

Generative Pre-Trained Transformer for Design Concept Generation: An Exploration

Abstract

Publication series

Conference

Research Keywords

Publisher's Copyright Statement

Access Output

Other Access Options

Permanent Link

Other Files and Links

Fingerprint

Cite this