Workshop 1

Should AI be More Like Shape Grammars? Machine Learning and Design

Monday 6 July 2026, 2:00 pm to 5:30 pm

Workshop Chairs
Athanassios Economou, Georgia Institute of Technology <thanos@gatech.edu >

Iestyn Jowers, Open University <iestyn.jowers@open.ac.uk>

Workshop Committee

Chris Earl, Open University

George Stiny, MIT

Alex Haridis, PHarvard University

Deborah Benros, University of East London

Thomas Wortmann, University of Stuttgart

Sotirios Kotsopoulos, Technical University of Athens 

Goals of the Workshop

In the paper “Artificial Intelligence” (https://doi.org/10.1068/b060353), published almost fifty years ago, James Gips, co-inventor (with George Stiny) of Shape Grammars, positioned art and design as  subfields of artificial intelligence, alongside computer vision, natural-language understanding, mathematics, medicine, games and more. Shape grammars formalize visual design, building on shapes and rules, and are uniquely placed to take on questions about design.  These range from conceptual enquiries into design generation, visual perception, evaluation and aesthetics as well as practical tools in computational design, CAD, visualization and making. Questions posed by Gips in the paper have been repeatedly asked by researchers over the decades, and are again being asked today in the context of the most recent AI revolution:

·      What procedures can be used to create designs?

·      What procedures can be used to understand and evaluate existing designs?

·      How can these procedures be specified as algorithms?

The fundamental insights provided by shape grammars to these questions have pointed to very different ways of computing, machine learning and design as well as to AI more generally.

Through short presentations and open discussion, the workshop aims to explore how shape grammar theory, methods and tools applied to design can provide insight and guidance for AI approaches to these tasks and AI more widely. The workshop welcomes open and speculative inputs from multiple disciplines including  AI, Design, Computing, Architecture and Aesthetics. 

Workshop Format

Through short presentations and open discussion, the workshop aims to explore how shape grammar theory, methods and tools applied to design can provide insight
and guidance for AI approaches to these tasks and AI more widely. The workshop welcomes open and speculative inputs from multiple disciplines including AI, Design, Computing, Architecture and Aesthetics. 

Themes for presentations and discussion include (but not confined to these):

1)      Generative design: rules and schemas - from shape to AI

2)      Visual perception: seeing and describing as a core AI capability

3)      Evaluation: design assessment in predefined and determinate categories vs emergent properties

4)      Aesthetics: value judgments in design and AI

5)      Computational design: implications of ambiguity e.g. in shape matching techniques, for AI computation

6)      Computer Aided Design (CAD): mining CAD descriptions for generating new designs

7)      Visualization: languages of shapes for AI

8)      Making: material transformations beyond visualisation

Workshop format

Introduction: 15 minutes
10 min presentations followed by 10 min discussion focused on presentation: 1-2 hours
Group discussion focused on workshop goals: 1-2 hours
Wrap up: 15 minutes

Submissions

Email workshop submissions as a pdf file to the Conference Chairs (thanos@gatech.edu,iestyn.jowers@open.ac.uk) with the subject DCC'26 Workshop 1.

Workshop Registration

All attendees at the workshop need to register either as an addition to the DCC'26 conference registration at a cost of €25 + VAT, or if not registered for the conference at a cost of €50 + VAT. Please go the DCC'26 Registration page to add this workshop to your registration.