Deep generative models of 3D shapes have received a great deal of research interest. Yet, almost all of them generate discrete shape representations, such as voxels, point clouds, and polygon meshes. We present the first 3D generative model for a drastically different shape representation--describing a shape as a sequence of computer-aided design (CAD) operations. Unlike meshes and point clouds, CAD models encode the user creation process of 3D shapes, widely used in numerous industrial and engineering design tasks. However, the sequential and irregular structure of CAD operations poses significant challenges for existing 3D generative models. Drawing an analogy between CAD operations and natural language, we propose a CAD generative network based on the Transformer. We demonstrate the performance of our model for both shape autoencoding and random shape generation. To train our network, we create a new CAD dataset consisting of 178,238 models and their CAD construction sequences. We have made this dataset publicly available to promote future research on this topic.
more »
« less
This content will become publicly available on March 1, 2025
A Study on Generative Design Reasoning and Students' Divergent and Convergent Thinking
Computer-aided design (CAD) is a standard design tool used in engineering practice and by students. CAD has become increasingly analytic and inventive in incorporating artificial intelligence (AI) approaches to design, e.g., generative design (GD), to help expand designers' divergent thinking. However, generative design technologies are relatively new, we know little about generative design thinking in students. This research aims to advance our understanding of the relationship between aspects of generative design thinking and traditional design thinking. This study was set in an introductory graphics and design course where student designers used Fusion 360 to optimize a bicycle wheel frame. We collected the following data from the sample: divergent and convergent psychological tests and an open-ended response to a generative design prompt (called the generative design reasoning elicitation problem). A Spearman's rank correlation showed no statistically significant relationship between generative design reasoning and divergent thinking. However, an analysis of variance found a significant difference in generative design reasoning and convergent thinking between groups with moderate GD reasoning and low GD reasoning. This study shows that new computational tools might present the same challenges to beginning designers as conventional tools. Instructors should be aware of informed design practices and encourage students to grow into informed designers by introducing them to new technology, such as generative design.
more » « less- Award ID(s):
- 1918847
- PAR ID:
- 10545395
- Publisher / Repository:
- American Society of Mechanical Engineers
- Date Published:
- Journal Name:
- Journal of Mechanical Design
- Volume:
- 146
- Issue:
- 3
- ISSN:
- 1050-0472
- Page Range / eLocation ID:
- 031405 (10 pages)
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Deep generative models of 3D shapes have received a great deal of research interest. Yet, almost all of them generate discrete shape representations, such as voxels, point clouds, and polygon meshes. We present the first 3D generative model for a drastically different shape representation—describing a shape as a sequence of computer-aided design (CAD) operations. Unlike meshes and point clouds, CAD models encode the user creation process of 3D shapes, widely used in numerous industrial and engineering design tasks. However, the sequential and irregular structure of CAD operations poses significant challenges for existing 3D generative models. Drawing an analogy between CAD operations and natural language, we propose a CAD generative network based on the Transformer. We demonstrate the performance of our model for both shape autoencoding and random shape generation. To train our network, we create a new CAD dataset consisting of 178,238 models and their CAD construction sequences. We have made this dataset publicly available to promote future research on this topic.more » « less
-
more » « less
Abstract In engineering design, engineers must be able to think creatively, effectively toggling between divergent thinking (developing multiple novel ideas) and convergent thinking (pursuing an appropriate idea using engineering analyses). However, creative thinking is not emphasized in many undergraduate engineering programs. In this empirical study, we analyze the divergent thinking of teams working on a virtual laboratory project. Fifteen student teams' solution paths–as represented by Model Maps–were analyzed to characterize and compare the various elements of divergent thinking: fluency, flexibility, and originality. The solution paths of these teams were compared in two physical laboratory projects and to experts completing the same virtual laboratory project. We found that students demonstrated more divergent thinking in the virtual laboratory project than in the physical laboratory projects; yet, divergent thinking and quality of solution did not correlate. There was little difference between measured elements of divergent thinking between student teams and experts.
-
Recent reforms in science education have supported the inclusion of engineering in K- 12 curricula. To this end, many science classrooms have incorporated engineering units that include design tasks. Design is an integral part of engineering and helps students think in creative and interdisciplinary ways. In this study, we examined middle-school students’ naturally occurring design conversations in small design teams and their learning of science as a result of engaging in an engineering and science unit. We found that the proportion of different thought processes used by boys and girls was quite similar. Both girls and boys produced a higher percentage of ideas or thoughts associated with divergent thinking, but a lower proportion in convergent thinking, evaluative thinking, and cognitive memory. In addition, gender composition of design teams influenced thought processes expressed by girls and boys. Interestingly, in mixed teams, both girls and boys expressed less divergent thinking than those in single-sex teams. With regard to science content learning, both girls and boys showed statistically significant learning gains. There were no significant gender differences in the pre- and post-test scores. These results suggest that participating in an engineering design task in small design teams provided students opportunities to engage in productive thinking and enhance their learning of the targeted science concept—ecosystems.more » « less
-
The research presented in this paper tested whether drawing concept maps changes how engineering students construct design problem statements and whether these differences are observable in their brains. The process of identifying and constructing problem statements is a critical step in engineering design. Concept mapping has the potential to expand the problem space that students explore through the attention given to the relationship between concepts. It helps integrate existing knowledge in new ways. Engineering students (n=66) were asked to construct a problem statement to improve mobility on campus. Half of these students were randomly chosen to first receive instructions about how to develop a concept map and were asked to draw a concept map about mobility systems on campus. The semantic similarity of concepts in the students’ problem statements, the length of their problem statements, and their neurocognition when developing their statements were measured. The results indicated that students who were asked to first draw concept maps produced a more diverse problem statement with less semantically similar words. The students who first developed concept maps also produce significantly longer problem statements. Concept mapping changed students’ neurocognition. The students who used concept mapping elicited less cognitive activation in their left prefrontal cortex (PFC) and more concentrated activation in their right PFC. The right PFC is generally associated with divergent thinking and the left PFC is generally associated with convergent and analytical thinking. These results provide new insight into how educational interventions, like concept mapping, can change students’ cognition and neurocognition. Better understanding how concept maps, and other tools, help students approach complex problems and the associated changes that occur in their brain can lay the groundwork for novel advances in engineering education that support new tools and pedagogy development for design.more » « less
