eSketch: A Multimodal Generative AI Design Interface and Human-Centered Approach for a Neural Design Phase
Research Team
Our Motivation:
"Current Generative Artificial Intelligence (GenAI) methodologies are not capable of generating Building Information Models (BIMs) directly. This limitation can hinder their effectiveness in providing comprehensive insights for informed investments in GenAI"
Research Contribution
Addressing the question of whether these GenAI methods enhance design skills, our research aims to develop eSketch, a design interface seamlessly integrated with GenAI techniques and human-centered approaches.
The primary goal of this development is to evaluate the overall user experience and its correlation with design skills. To achieve this, we will conduct comparative user studies, comparing the performance of our methods with established benchmarks.
For instance, we will assess how eSketch users' design skills compare to those using traditional design tools and workflows.
Problem
Practical Problem
In the context of Industrialized Construction (IC), the early-stage designs frequently lack essential constructability insights. This deficiency results in delays, as modular construction companies invest weeks in the process of modifying designs to align with their specific building modules and construction methodologies. Consequently, latency emerges between the schematic design phase and the development stages.
Solution
The solution consists in the development of an interface for evaluating the GenAI methodology in the context of Industrialized Construction (IC) project design.
Added Value For The Industry
It will empower industry partners to generate Building Information Models (BIM) from initial design concepts, thereby reducing the time lag between the schematic and development phases within the Industrialized Construction framework.
Cooperation Partner
Timeline
Date | Activity | Outcome |
October 2023 | Research became awarded |
|
April 2024 | Finalize an initial working deep generative model |
|
June 2024 | Initial Prototype |
|
Summer 2024 | Initial User Interface Studies |
|
If you want to participate in the project please reach out to Alberto Tono.