No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
How Long Until We Are (Psychologically) Safe? A Longitudinal Investigation of Psychological Safety in Virtual Engineering Design Teams in Education
Abstract
This paper investigates team psychological safety (N = 34 teams) in a synchronous online engineering design class spanning 4 weeks. While work in this field has suggested that psychological safety in virtual teams can facilitate knowledge-sharing, trust among teams, and overall performance, there have been limited investigations of the longitudinal trajectory of psychological safety, when the construct stabilizes in a virtual environment, and what factors impact the building of psychological safety in virtual teams.
ResearchGate has not been able to resolve any citations for this publication.
Shape grammar interpreters have been studied for more than forty years addressing several areas of design research including architectural, engineering, and product design. At the core of all these implementations, the operation of embedding – the ability of a shape grammar interpreter to search for subshapes in a geometry model even if they are not explicitly encoded in the database of the system – resists a general solution. It is suggested here that beyond a seemingly long list of technological hurdles, the implementation of shape embedding, that is, the implementation of the mathematical concept of the “part relation” between two shapes, or equivalently, between two drawings, or between a shape and a design, is the single major obstacle to take on.
This research identifies five challenges underlying the implementation of shape embedding and shape grammar interpreters at large: 1) complex entanglement of the calculations required for shape embedding and a shape grammar interpreter at large, with those required by a CAD system for modeling and modifying geometry; 2) accumulated errors caused by the modeling processes of CAD systems; 3) accumulated errors caused by the complex calculations required for the derivation of affine, and mostly, perspectival transformations; 4) limited support for indeterminate shape embedding; 5) low performance of the current shape embedding algorithms for models consisting of a large number of shapes.
The dissertation aims to provide a comprehensive engineering solution to all these five challenges above. More specifically, the five contributions of the dissertation are: 1) a new architecture to separate the calculations required for the shape embedding and replacement (appropriately called here Shape Machine) vs. the calculations required by a CAD system for the selection, instantiation, transformation, and combination of shapes in CAD modeling; 2) a new modeling calibration system to ensure the effective translation of geometrical types of shapes to their maximal representations without cumulative calculating errors; 3) a new dual-mode system of the derivation of transformations for shape embedding, including a geometric approach next to the known algebraic one, to implement the shape embedding relation under the full spectrum of linear transformations without the accumulated errors caused by the current algorithms; 4) a new multi-step mechanism that resolves all cases of indeterminate embeddings for shapes having fewer registration points than those required for a shape embedding under a particular type of transformation; and 5) a new data representation for hyperplane intersections, the registration point signature, to allow for the effective calculation of shape embeddings for complex drawings consisting of a large number of shapes. All modules are integrated into a common computational framework to test the model for a particular type of shapes – the shapes consisting of lines in the Euclidean plane in the algebra U_12.
The role of using performance-based design optimization in early-stage architectural design to prevent poorly designed buildings has been increasingly recognized by researchers and designers. Recently, a large amount of research has been made focusing on technical advancement, which, however, also reflects the limited research on how this technique can be applied to the design process and how it can aid designers when confronting ill-defined design problems. In this regard, this paper centers on the design approaches assisted by using performance-based design optimization. The paper proposes two optimization-aided design approaches that were identified in the previous applications of a design tool, called EvoMass, and showcases these approaches in case-study designs. The case studies demonstrate how the use of performance-based design optimization can facilitate designers’ reflection and exploration. With the demonstration of the design approaches, we discuss the utility of performance-based design optimization in assisting architects in the early design stages.
We replicate a design ideation experiment (Goucher-Lambert et al., 2019) with and without inspirational stimuli and extend data collection sources to eye-tracking and a think aloud protocol to provide new insights into generated ideas. Preliminary results corroborate original findings: inspirational stimuli have an effect on idea output and questionnaire ratings. Near and far inspirational stimuli increased participants’ idea fluency over time and were rated more useful than control. We further enable experiment reproducibility and provide publicly available data.
Over the past few decades, architectural practice and, consequently, the design studio have been increasingly challenged. Indeed, the development of digital tools and parametric design, in particular, has given rise to a new type of architectural knowledge. Among the IJAC publications over the past three years, we highlight the current diversity of vocabulary used to discuss this knowledge and develop why we focus our study on conceptual knowledge. We then report a learning situation through studio design education. This paper finally presents the steps developed to measure this knowledge and hypothesizes on the future work needed in order to have relevant quantitative results. The purpose of this paper is to observe the evolution of students' understanding when shifting from a traditional teacher-student relationship to an engaging learning environment, considering the specificities of parametric, and not to suggest a strict method to follow when learning parametric. This could guide teachers to adapt to their own situations.
This paper presents a ‘skeletal’ parametric schema to generate residential building layout configurations for performance-based design optimization. The schema generates residential building layout configurations using a set of ‘skeletal’ lines that are created based on various design elements and coincident with factors such as walkways, spacing, and setback requirements. As such, the schema is able to generate diverse and legitimate design alternatives. With the proposed parametric schema, a case-study optimization is carried out for a Singapore Housing Development Board (HDB) project. The case study considers a set of performance criteria and produces results with higher practical referential value. The case study demonstrates that the optimization with the parametric schema can improve the overall quality of the design and provide designers with various design options.
The role of optimization-based design exploration in early-stage architectural design has been increasingly recognized and valued. It has been widely considered an effective approach to achieving performance-informed and performance-driven design. Nevertheless, there is little research into how such design exploration can be adapted to various early-stage architectural design tasks. With this motivation, this paper revolves around a computer-aided design workflow for early-stage building massing design optimization and exploration while presenting three workshop case studies to demonstrate how the workflow can be intertwined with the design process. The design workflow is based on EvoMass, an integrated building massing design generation and optimization tool in Rhino-Grasshopper. The case study illustrates task-specific applications of the design workflow for synthesizing building design, finding design precedents, and understanding the interrelationship between formal attributes and building performance. The paper concludes by discussing the relevant efficacy of the design workflow for architectural design.
Formal computational approaches in the realm of engineering and architecture, such as parametric modelling and optimization, are increasingly powerful, allowing for systematic and rigorous design processes. However, these methods often bring a steep learning curve, require previous expertise, or are unintuitive and unnatural to human design. On the other hand, analog design methods such as hand sketching are commonly used by architects and engineers alike, and constitute quick, easy, and almost primal modes of generating and transferring design concepts, which in turn facilitates the sharing of ideas and feedback. In the advent of increasing computational power and developments in data analysis, deep learning, and other emerging technologies, there is a potential to bridge the gap between these seemingly divergent processes to develop new hybrid approaches to design. Such methods can provide designers with new opportunities to harness the systematic and data-driven power of computation and performance analysis while maintaining a more creative and intuitive design interface. This paper presents a new method for interpreting human designs in sketch format and predicting their structural performance using recent advances in deep learning. The paper also demonstrates how this new technique can be used in design workflows including performance-based guidance and interpolations between concepts.
Shape queries based on shape embedding under a given Euclidean, affine, or linear transformation are absent from current CAD systems. The only systems that have attempted to implement shape embedding are the shape grammar interpreters albeit with promising but inconclusive results. The work here identifies all possible 14 cases of shape embedding with respect to the number of available registration points, four for determinate cases and ten for indeterminate ones, and an approach is sketched to take on the complexities underlying the indeterminate cases. All visual calculations are done with shapes consisting of straight lines in the Euclidean plane within the algebra U ij for i = 1 the dimension of lines and j = 2 the dimension of space in which the lines are defined, transformed and combined. Aspects of interface design and integration to current work design workflows are deliberately left aside.
Shape grammar interpreters have been studied for more than forty years addressing several areas of design research including architectural, engineering, and product design. At the core of all these implementations, the operation of embedding—the ability of a shape grammar interpreter to search for subshapes in a geometry model even if they are not explicitly encoded in the database of the system—resists a general solution. Here, a detailed account on various constructions of embedding is provided, including determinate and indeterminate ones, to give a sense of the rising complexity of their implementation in a shape grammar interpreter, and to provide a visual map of the work accomplished in the field so far, and the work ahead too.
https://www.tandfonline.com/doi/full/10.1080/21650349.2022.2021480
Artificial intelligence (AI) has shown its promise in assisting human decision-making. However, humans' inappropriate decision to accept or reject suggestions from AI can lead to severe consequences in high-stakes AI-assisted decision-making scenarios. This problem persists due to insufficient understanding of human trust in AI. Therefore, this research studies how two types of human confidence that affect trust, their confidence in AI and confidence in themselves, evolve and affect humans’ decisions. A cognitive study and a quantitative model together examine how changing positive and negative experiences affect these confidences and ultimate decisions. Results show that human self-confidence, not their confidence in AI, directs the decision to accept or reject AI suggestions. Furthermore, this work finds that humans often misattribute blame to themselves and enter a vicious cycle of relying on a poorly performing AI. Findings reveal the need and provide insights to effectively calibrate human self-confidence for successful AI-assisted decision-making.
This paper presents an extended evaluation of tensor-based representations of graph-based architectural room configurations. This experiment is a continuation of examination of recognition of semantic architectural features by contemporary standard deep learning methods. The main aim of this evaluation is to investigate how the deep learning models trained using the relation tensors as data representation means perform on data not available in the training dataset. Using a straightforward classification task, stepwise modifications of the original training dataset and manually created spatial configurations were fed into the models to measure their prediction quality. We hypothesized that the modifications that influence the class label will not decrease this quality, however, this was not confirmed and most likely the latent non-class defining features make up the class for the model. Under specific circumstances, the prediction quality still remained high for the winning relation tensor type.
In the past two decades, there has been increasing use of semantic networks in engineering design for supporting various activities, such as knowledge extraction, prior art search, idea generation and evaluation. Leveraging large-scale pre-trained graph knowledge databases to support engineering design-related natural language processing (NLP) tasks has attracted a growing interest in the engineering design research community. Therefore, this paper aims to provide a survey of the state-of-the-art semantic networks for engineering design and propositions of future research to build and utilize large-scale semantic networks as knowledge bases to support engineering design research and practice. The survey shows that WordNet, ConceptNet and other semantic networks, which contain common-sense knowledge or are trained on non-engineering data sources, are primarily used by engineering design researchers to develop methods and tools. Meanwhile, there are emerging efforts in constructing engineering and technical-contextualized semantic network databases, such as B-Link and TechNet, through retrieving data from technical data sources and employing unsupervised machine learning approaches. On this basis, we recommend six strategic future research directions to advance the development and uses of large-scale semantic networks for artificial intelligence applications in engineering design.
Virtual Personal Assistants (VPAs) are becoming so widely available that considerations are being made as to whether to begin including them in self-driving vehicles. While research has been done exploring human interactions with single VPAs, there has been little work exploring human interactions and human mental models with interconnected systems. As companies like Amazon consider whether to integrate Alexa in their self-driving car, research needs to be done to explore whether individual's mental model of these systems is of a single system or if every embodiment of the VPA (e.g., echo) represents a different VPA. Knowing this will allow researchers and practitioners to apply existing models of trust, and predict whether high trust in the Siri that exists in an iPhone will carry over into high (and potentially miscalibrated) trust in Apple's Siri-directed self-driving vehicle. Results indicate that there is not one consistent mental model that users have, and provides the framework for greater exploration into individual differences and the determinants that affect users' mental model.
Combinatorial Design such as configuration design, design optioneering, component selection, and generative design, is common across engineering. Generating solutions for a combinatorial design task often involves the application of classical computing solvers that can either map or navigate design spaces. However, it has been observed that classical computing resource power-law scales with many design space models. This observation suggests classical computing may not be capable of modelling our future design space needs.
To meet future design space modelling needs, this paper examines quantum computing and the characteristics that enables its resources to scale polynomially with design space size. The paper then continues to present a combinatorial design problem that is subsequently represented, constrained and solved by quantum computing. The results of which are the derivation of an initial set of circuits that represent design space constraints. The study shows the game-changing possibilities of quantum computing as an engineering design tool and is the start of an exciting new journey for design research.
The representation of the product use context is a well established design practice in Engineering Design. Recently, design theory is studying the product interaction involving several cognitive aspects such as the possible conditions in which a wrong interaction occurs. The aim of this paper is to find a quantitative evidence of the causes of these misuses. In particular, this study focuses on the detection of bad design and biases.
In this paper, we propose a method that helps to the automatic detection of bad design and biases from patents. The method is based on an approach that defines syntactic rules to detect sentences containing these artifacts. These rules are defined based on an exploratory analysis of the explicit mention of “bad design” and “bias” and then, tested with multiple experiments on a sample of patents. The results give a first quantitative evidence of the presence of bad design and biases in patents and consequently of their importance in the design theory. In particular, it is provided a fine grain analysis of the linguistic structure of sentences containing these artifacts helping designers in detecting automatically them from patents.
The objective evaluation of empirical studies is an important part when assessing demand and validating design methods. However, metrics that can also map cognitive processes during design are still lacking. In order to address this problem, an online study with 12 participants was conducted. The aim of this investigation was to find a relation between cognitive load and performance in engineering design tasks. To assess the cognitive load, the NASA-RTLX questionnaire was used as an established measurement tool and was related to the results achieved by the participants. The results show that there is a correlation between the two investigated parameters. Based on a statistical analysis a correlation between increasing cognitive load and a decrease in performance could be identified. The tasks used produce comparable results to other studies investigating cognitive load, but the task causing the highest cognitive load shows the widest scatter in performance. The u-curve as suggested by the state of the art was not visible in the study’s results, but the cognitive load should be nevertheless used for studies of design processes, because it may reveal a need for methodical support.
This paper presents an EEG (Electroencephalography) study that explores correlations between the neurophysiological activations, the nature of the design task and its outputs. We propose an experimental protocol that covers several design-related tasks: including fundamental activities (e.g. idea generation and problem-solving) as well as more comprehensive task requiring the complex higher-level reasoning of designing. We clustered the collected data according to the characteristics of the design outcome and measured EEG alpha band activation during elementary and higher-level design task, whereas just the former yielded statistically significant different behaviour in the left frontal and occipital area. We also found a significant correlation between the ratings for elementary sketching task outcomes and EEG activation at the higher-level design task. These results suggested that EEG activation enables distinguishing groups according to their performance only for elementary tasks. However, this also suggests a potential application of EEG data on the elementary tasks to distinguish the designers' brain response during higher-level of design task.
This paper asked participants to assess four selected expert-rated Taiwan International Student Design Competition (TISDC) products using four methods: Consensual Assessment Technique (CAT), Creative Product Semantic Scale (CPSS), Product Creativity Measurement Instrument (PCMI), and revised Creative Solution Diagnosis Scale (rCSDS). The results revealed that, between experts and non-experts, the ranking results by the CAT and CPSS were the same, while the ranking results of the rCSDS were different. The CAT, CPSS, and TISDC methods provided the same results indicating that raters may return the same results on creativity assessment, and the results are not affected by the selected methods.
If it is necessary to use non-experts to assess creativity and the creativity results are expected to be the same with that of experts, asking non-expert raters to use CPSS to assess creativity and then ranking the creativity score is more reliable. The study offers a contribution to the creativity domain on deciding which methods may be more reliable from a comparison perspective.
Whilst prior works have characterised the affordances of prototyping methods in terms of generating knowledge about a product or process, the types, or ‘dimensions’ of knowledge towards which they contribute are not fully understood. In this paper we adapt the concept of ‘design domains’ as a method to interpret, and better understand the contributions of different prototyping methods to design knowledge in new product development. We first synthesise a set of ten dimensions for design knowledge from a review of literature in design-related fields. A study was then conducted in which participants from engineering backgrounds completed a Likert-type questionnaire to quantify the perceived contributions to design knowledge of 90 common prototyping methods against each dimension. We statistically analyse results to identify patterns in the knowledge contribution of different methods. Results reveal that methods exhibit significantly different contribution profiles, suggesting different methods to be suited to different knowledge. Thus, this paper indicates potential for new methods, methodology and processes to leverage such characterisations for better selection and sequencing of methods in the prototyping process.
Hackathons are short design events at which participants collaboratively progress through the entire design process. They pose opportunities for design research, but the existing research is limited, as is the understanding of design activity at hackathons. In our study, we summarize the hackathon design process of 10 interview participants from varying disciplines, levels of experience, and hackathon events. The summarized account reveals a decreased emphasis on the beginning phases of the design process, mainly problem definition, but an increased emphasis on the end, specifically the pitch portion of the event. These differences are mainly due to the limited time frame. We further assess the effect of time limitations at hackathons by comparing hackathons to other instances of design, emphasizing the impact of time constraints on iteration. We conclude our discussion with an exploration of the role expertise has on the design process by comparing the accounts of designers and developers.
In this paper, we present results from an experiment using EEG to measure brain activity and explore EEG frequency power associated with gender differences of professional industrial designers while performing two prototypical stages of constrained and open design tasks, problem-solving and design sketching. Results indicate no main effect of gender. However, among other main effects, a consistent main effect of hemisphere for the six frequency bands under analysis was found. In the problem-solving stage, male designers show higher alpha and beta bands in channels of the prefrontal cortices and female designers in the right occipitotemporal cortex and secondary visual cortices. In the design sketching stage, male designers show higher alpha and beta bands in the right prefrontal cortex, and female designers in the right temporal cortex and left prefrontal cortex, where higher theta is also found. Prioritising different cognitive functions seem to play a role in each gender's approach to constrained and open design tasks. Results can be useful to design professionals, students and design educators, and for the development of methodological approaches in design research and education.
Engineers often need to discover and learn designs from unfamiliar domains for inspiration or other particular uses. However, the complexity of the technical design descriptions and the unfamiliarity to the domain make it hard for engineers to comprehend the function, behavior, and structure of a design. To help engineers quickly understand a complex technical design description new to them, one approach is to represent it as a network graph of the design-related entities and their relations as an abstract summary of the design. While graph or network visualizations are widely adopted in the engineering design literature, the challenge remains in retrieving the design entities and deriving their relations. In this paper, we propose a network mapping method that is powered by Technology Semantic Network (TechNet). Through a case study, we showcase how TechNet’s unique characteristic of being trained on a large technology-related data source advantages itself over common-sense knowledge bases, such as WordNet and ConceptNet, for design knowledge representation.
Decisions made by human-AI teams (e.g., AI-advised humans) are increasingly common in high-stakes domains such as healthcare, criminal justice, and finance. Achieving high team performance depends on more than just the accuracy of the AI system: Since the human and the AI may have different expertise, the highest team performance is often reached when they both know how and when to complement one another. We focus on a factor that is crucial to supporting such complementary: the human’s mental model of the AI capabilities, specifically the AI system’s error boundary (i.e. knowing “When does the AI err?”). Awareness of this lets the human decide when to accept or override the AI’s recommendation. We highlight two key properties of an AI’s error boundary, parsimony and stochasticity, and a property of the task, dimensionality. We show experimentally how these properties affect humans’ mental models of AI capabilities and the resulting team performance. We connect our evaluations to related work and propose goals, beyond accuracy, that merit consideration during model selection and optimization to improve overall human-AI team performance.
Automated design synthesis has the potential to revolutionize the modern engineering design process and improve access to highly optimized and customized products across countless industries. Successfully adapting generative machine learning to design engineering may enable such automated design synthesis and is a research subject of great importance. We present a review and analysis of deep generative machine learning models in engineering design. Deep generative models (DGMs) typically leverage deep networks to learn from an input dataset and synthesize new designs. Recently, DGMs such as feedforward neural networks (NNs), generative adversarial networks (GANs), variational autoencoders (VAEs), and certain deep reinforcement learning (DRL) frameworks have shown promising results in design applications like structural optimization, materials design, and shape synthesis. The prevalence of DGMs in engineering design has skyrocketed since 2016. Anticipating the continued growth, we conduct a review of recent advances to benefit researchers interested in DGMs for design. We structure our review as an exposition of the algorithms, datasets, representation methods, and applications commonly used in the current literature. In particular, we discuss key works that have introduced new techniques and methods in DGMs, successfully applied DGMs to a design-related domain, or directly supported the development of DGMs through datasets or auxiliary methods. We further identify key challenges and limitations currently seen in DGMs across design fields, such as design creativity, handling constraints and objectives, and modeling both form and functional performance simultaneously. In our discussion, we identify possible solution pathways as key areas on which to target the future work.
During the early stages of design exploration, competing design strategies are typically considered. This chapter presents a design method, supported by a novel type of evolutionary algorithm, that maintains a heterogeneous population of design variants based on competing design strategies. Each strategy defines its own search space of design variants, all sharing a common generative concept or idea. A population of design variants is evolved through a process of selection and variation. As evolution progresses, some design strategies will become extinct while others will gradually dominate the population. A demonstration is presented showing how a designer can explore competing strategies by running a series of iterative evolutionary searches. The evolutionary algorithm has been implemented on a cloud platform, thereby allowing populations design variants to be processed in parallel. This results in a significant reduction in computation time, allowing thousands of designs to be evolved in just a few minutes.
Generative design systems can generate a wide panoply of solutions, from which designers search for those that best suit their interests. However, without guidance, this search can be highly inefficient, and many interesting solutions may remain unexplored. This problem is mitigated with automated exploration methods. Still, the ones typically provided by generative design tools are mostly based on black-box methods that drastically reduce the role of the designer, while more straightforward white-box mechanisms are dispersedly found in specific applications. This paper proposes the Navigator tool, which gathers a set of white-box mechanisms that automate the generation of default, random, similar and hybrid designs and design subspaces, while also supporting the generation of design collections. The proposed mechanisms were tested with two generative systems that create, respectively, tower and chair designs. We expect that, by providing understandable mechanisms for navigating design spaces, designers can become more engaged in the search process.
This paper reexamines Uij algebras which constitute the framework for calculations with shapes in the context of shape grammars. The calculations are done in a Boolean fashion and may have both spatial and nonspatial components. We start with partial algebras of geometric elements, where calculations are purely spatial, and gradually introduce nonspatial ones to build shapes and their standard algebras (Uij). Based on this Uij is decomposed into several algebras that do mostly spatial calculations the result of which is then turned into shape in a nonspatial fashion. The resulting decomposition algebra does the same job as Uij while keeping the spatial and nonspatial computations as separate as possible. Besides being a novel structure, the new algebra allows for more efficient computer applications and its version supports simultaneous calculations with shapes and their boundaries.
Common tests of spatial skills do not simply test one’s ability to mentally manipulate shapes. Instead, many popular assessments depend on a separate ability to comprehend two-dimensional graphical depictions of three-dimensional objects. Two categories of evidence are presented: 1) a discussion of the visual problems present in the stimuli commonly used in spatial skills tests, and 2) a critical review of studies which have shown improved performance on spatial skills tests by making the stimuli either clearer or more realistic. We conclude that the graphical interpretation factor is likely an example of construct-irrelevant variance which may reduce the validity of spatial skills assessments and introduce bias in favor of individuals with past experience with a particular style of engineering graphics or individuals who leverage certain problem-solving strategies.
Functional magnetic resonance imaging (fMRI) enables identification of the brain regions and networks underpinning cognitive tasks. It has the potential to significantly advance cognitive design science, but is challenging to apply in design studies and methodological guidance for design researchers is lacking. In this Research Note, we reflect on our experiences and other work to outline the activities involved in developing and executing fMRI design studies. The implications for research quality at each stage are highlighted. We then consider the challenges for fMRI research on design and make recommendations for addressing them. Four critical areas are identified: establishing experimental protocols; establishing a cognitive design ontology; generating foundational knowledge about brain activation; and balancing fMRI constraints against ecological validity.
Artificial Intelligence (AI) has had a strong presence in engineering design for decades, and while theory, methods, and tools for engineering design have advanced significantly during this time, many grand challenges remain. Modern advancements in AI, including new strategies for capturing, storing, and analyzing data, have the potential to revolutionize engineering design processes in a variety of ways. The purpose of this special issue is to consolidate recent research activities that utilize existing or new AI methods to advance engineering design knowledge and capabilities.
This paper presents a framework to describe and explain human-machine collaborative design focusing on Design Space Exploration (DSE), which is a popular method used in the early design of complex systems with roots in the well-known design as exploration paradigm. The human designer and a cognitive design assistant are both modeled as intelligent agents, with an internal state (e.g., motivation, cognitive workload), a knowledge state (separated in domain, design process, and problem specific knowledge), an estimated state of the world (i.e., status of the design task) and of the other agent, a hierarchy of goals (short-term and long-term, design and learning goals) and a set of long-term attributes (e.g., Kirton’s Adaption-Innovation inventory style, risk aversion). The framework emphasizes the relation between design goals and learning goals in DSE, as previously highlighted in the literature (e.g., Concept-Knowledge theory, LinD model) and builds upon the theory of common ground from human-computer interaction (e.g., shared goals, plans, attention) as a building block to develop successful assistants and interactions. Recent studies in human-AI collaborative DSE are reviewed from the lens of the proposed framework, and some new research questions are identified. This framework can help advance the theory of human-AI collaborative design by helping design researchers build promising hypotheses, and design studies to test these hypotheses that consider most relevant factors.
Final concepts are often not the most creative or innovative design within the solution space. The purpose of this research is to gain insight into the decisions made in concept selection. In particular, we studied how designers link multiple decision-making elements together, including: actions (what people do), reasoning (why they do it), and design outcomes (an objective measure of engineering performance). Fifty-seven participants were tasked with solving a design challenge relating to a robotic gripper by selecting a design within a predefined design space. Each design had a corresponding measure (termed “success rate”) which enabled each designer’s performance to be quantified and compared against other designers. The task was hosted on an interactive interface in which design actions were collected. A post-task survey probed for the reasoning behind design actions. Characterization of decision-making behavior and reasoning was rooted in prior design literature. Design actions were quantified concerning the degree of design space explored and the decision-making strategies employed. Key results include design strategies such as manipulation techniques, the impact of maximum observed success rates, and a willingness to submit an alternative solution which influenced design outcomes. Although designer preferences validated the design strategies identified, there was no correlation between the decision factors considered and improved outcomes. The methods and findings from this work assessed the underlying dynamics when engineers selected less innovative or creative solutions and recommended decision-making strategies that should be considered to improve design outcomes.
The Theory of Inventive Problem Solving (TRIZ) method and toolkit provides a well-structured approach to support engineering design with pre-defined steps: interpret and define the problem, search for standard engineering parameters, search for inventive principles to adapt, and generate final solutions. The research presented in this paper explores the neuro-cognitive differences of each of these steps. We measured the neuro-cognitive activation in the prefrontal cortex (PFC) of 30 engineering students. Neuro-cognitive activation was recorded while students completed an engineering design task. The results show a varying activation pattern. When interpreting and defining the problem, higher activation is found in the left PFC, generally associated with goal directed planning and making analytical. Neuro-cognitive activation shifts to the right PFC during the search process, a region usually involved in exploring the problem space. During solution generation more activation occurs in the medial PFC, a region generally related to making associations. The findings offer new insights and evidence explaining the dynamic neuro-cognitive activations when using TRIZ in engineering design.
One of the most influential descriptions of design activity emphasizes how problems and solutions “co-evolve.” This concept has somehow escaped critical review and cross-disciplinary comparison, resulting in a fragmented approach to the subject. Reviewing the published literature on design co-evolution reveals that the term is used to refer to a range of distinct concepts, and the study of co-evolution has generated a number of elaborations and alternatives. Reviewing the broader literature in design and other disciplines further reveals that discussions of design co-evolution are disconnected from the history of relevant concepts in design research, and disconnected from a range of relevant concepts in other disciplines that describe creative work. Here I examine what the different concepts of design co-evolution are, how they have been modified and what they are related to. This leads to questioning the distinction between problems and solutions, defining them in relative terms, and drawing a connection between design co-evolution and design fixation.
It is well known that greenery biophilic design can improve health and productivity, but studies are still needed to quantify greenery dose and corresponding well-being benefits to support design practice. In this study, we investigated impacts of various greenery dose on workplace well-being from the perspectives of physiological, psychological, and productivity performance. A repeated measures experiment was conducted, in which green coverage ratios, 0, 0.2%, 5%, 12% and 20% were tested, and both health and productivity performance of 15 participants were measured by advanced devices, questionnaire, and tasks. Eye tracking tests were also recruited to verify that health and productivity benefits were caused by visual contact with greenery that attracted participants’ visual attention. Regarding productivity, we observed improving working performance at all the cases with greenery. But green coverage ratio 0.2% and 5% are not enough to create restorative effects to change psychological response and 20% ratio is too overwhelming. Meanwhile, only green coverage ratios 12% and 20% could cause the positive change on physiological brain activities. Therefore, we identified 12% green coverage ratio as the optimal greenery dose for the office after integrating the results on psychological, physiological, and productivity performance. This research outputs provide not only the important measurement datasets for quantitative biophilic design research, but also important understandings to support biophilic design practice since green coverage ratio is one of the most important design parameters.
Generative design problems often encompass complex action spaces that may be divergent over time, contain state-dependent constraints, or involve hybrid (discrete and continuous) domains. To address those challenges, this work introduces Design Strategy Network (DSN), a data-driven deep hierarchical framework that can learn strategies over these arbitrary complex action spaces. The hierarchical architecture decomposes every action decision into first predicting a preferred spatial region in the design space and then outputting a probability distribution over a set of possible actions from that region. This framework comprises a convolutional encoder to work with image-based design state representations, a multi-layer perceptron to predict a spatial region, and a weight-sharing network to generate a probability distribution over unordered set-based inputs of feasible actions. Applied to a truss design study, the framework learns to predict the actions of human designers in the study, capturing their truss generation strategies in the process. Results show that DSNs significantly outperform non-hierarchical methods of policy representation, demonstrating their superiority in complex action space problems.
Creativity assessments should be valid, reliable, and scalable to support various stakeholders (e.g., policy-makers, educators, corporations, and the general public) in their decision-making processes. Established initiatives toward scalable creativity assessments have relied on well-studied standardized tests. Although robust in many ways, most of these tests adopt unnatural and unmotivating environments for expression of creativity, mainly observe coarse-grained snippets of the creative process, and rely on subjective, resource-intensive, human-expert evaluations. This article presents a literature review of game-based creativity assessment and discusses how digital games can potentially address the limitations of traditional testing. Based on an original sample of 127 papers, this article contributes an in-depth review of 16 papers on 11 digital creativity assessment games. Despite the relatively small sample, a wide variety of design decisions are covered. Major findings and recommendations include identifying (1) a disconnect between the potential of scaling up assessment of creativity with the use of digital games, and the actual reach achieved in the examined studies (2) the need for complementary methods such as stealth assessment, algorithmic support and crowdsourcing when designing creativity assessment games, and (3) a need for interdisciplinary dialogs to produce, validate and implement creativity assessment games at scale.
Design concept evaluation is a key process in the new product development process with a significant impact on the product's success and total cost over its life cycle. This paper is motivated by two limitations of the state-of-the-art in concept evaluation: (1) The amount and diversity of user feedback and insights utilized by existing concept evaluation methods such as quality function deployment are limited. (2) Subjective concept evaluation methods require significant manual effort which in turn may limit the number of concepts considered for evaluation. A Deep Multimodal Design Evaluation (DMDE) model is proposed in this paper to bridge these gaps by providing designers with an accurate and scalable prediction of new concepts' overall and attribute-level desirability based on large-scale user reviews on existing designs. The attribute-level sentiment intensities of users are first extracted and aggregated from online reviews. A multimodal deep regression model is then developed to predict the overall and attribute-level sentiment values based on the features extracted from orthographic product images via a fine-tuned ResNet-50 model and from product descriptions via a fine-tuned BERT model, and aggregated using a novel self-attention-based fusion model. The DMDE model adds a data-driven, user-centered loop within the concept development process to better inform the concept evaluation process. Numerical experiments on a large dataset from an online footwear store indicate a promising performance by the DMDE model with 0.001 MSE loss and over 99.1% accuracy.
Ideation is a key phase in engineering design and brainstorming is an established method for ideation. A limitation of the brainstorming process is idea production tends to peak at the beginning and quickly decreases with time. In this exploratory study, we tested an innovative technique to sustain ideation by providing designers feedback about their neurocognition. We used a neuroimaging technique (fNIRS) to monitor students’ neurocognitive activations during a brainstorming task. Half received real-time feedback about their neurocognitive activation in their prefrontal cortex, a brain region associated with working memory and cognitive flexibility. Students who received the neurocognitive feedback maintained higher cortical activation and longer sustained peak activation. Students receiving the neurocognitive feedback demonstrated a higher percentage of right-hemispheric dominance, a region associated to creative processing, compared to the students without neurocognitive feedback. The increase in right-hemispheric dominance positively correlated with an increase in the number of solutions during concept generation and a higher design idea fluency. These results demonstrate the prospective use of neurocognitive feedback to sustain the cognitive activations necessary for idea generation during brainstorming. Future research should explore the effect of neurocognitive feedback with a more robust sample of designers and compare neurocognitive feedback with other types of interventions to sustain ideation.
We propose a large scalable engineering knowledge base as an integrated knowledge graph, comprising sets of (entity, relationship, entity) triples that are real-world engineering ‘facts’ found in the patent database. We apply a set of rules based on the syntactic and lexical properties of claims in a patent document to extract entities and their associated relationships that are supposedly meaningful from an engineering design perspective. Such a knowledge base is expected to support inferencing, reasoning, recalling in various engineering design tasks. The knowledge base has a greater size and coverage in comparison with the previously used knowledge bases in the engineering design literature.
This research aims to augment human cognition through the advancement and automation of mindmapping technologies, which could later support human creativity and virtual collaboration. Mindmapping is a visual brainstorming technique that allows problem solvers to utilize the human brain's ability to retrieve knowledge through similarity and association. While it is a powerful tool to generate concepts in any phase of problem-solving or design, the content of mindmaps is usually manually generated while listening or conversing and generating ideas, requiring a high cognitive load. This work introduces the development of a speech-driven automated mindmapping technology, called Speech2Mindmap. The specifics of the Speech2Mindmap algorithm are detailed, along with two case studies that serve to test its accuracy in comparison to human generated mindmaps, using audio recorded speech data as input. In the first case study, the Speech2Mindmap algorithm was evaluated on how well it represents manually generated human mindmapping output. The second case study evaluated the reliability of the Speech2Mindmap algorithm and examined the best performing methods and conditions to achieve the greatest similarity to human generated mindmaps. This research demonstrates that the Speech2Mindmap algorithm is capable of representing manually generated human mindmapping output, and found the best performing methods and conditions to generate a mindmap that is 80% similar, on average, to human generated mindmaps.
This paper introduces the Simulated Jet Engine Bracket Dataset (SimJEB) [WBM21]: a new, public collection of crowdsourced mechanical brackets and accompanying structural simulations. SimJEB is applicable to a wide range of geometry processing tasks; the complexity of the shapes in SimJEB offer a challenge to automated geometry cleaning and meshing, while categorical labels and structural simulations facilitate classification and regression (i.e. engineering surrogate modeling). In contrast to existing shape collections, SimJEB's models are all designed for the same engineering function and thus have consistent structural loads and support conditions. On the other hand, SimJEB models are more complex, diverse, and realistic than the synthetically generated datasets commonly used in parametric surrogate model evaluation. The designs in SimJEB were derived from submissions to the GrabCAD Jet Engine Bracket Challenge: an open engineering design competition with over 700 hand‐designed CAD entries from 320 designers representing 56 countries. Each model has been cleaned, categorized, meshed, and simulated with finite element analysis according to the original competition specifications. The result is a collection of 381 diverse, high‐quality and application‐focused designs for advancing geometric deep learning, engineering surrogate modeling, automated cleaning and related geometry processing tasks.
This research explores the effect of the structuredness of design concept generation techniques on temporal network neurocognition. Engineering graduate students (n = 30) completed three concept generation tasks using techniques with different levels of structuredness: brainstorming, morphological analysis, and TRIZ. Students’ brain activation in their prefrontal cortex (PFC) was measured using functional near-infrared spectroscopy (fNIRS). The temporal dynamic of central regions in brain networks were compared between tasks. Central regions facilitate functional interaction and imply information flow through the brain. A consistent central region appears in the medial PFC. Consistent network connections occurred across both hemispheres suggesting a concurrent dual processing of divergent and convergent thinking. This study offers novel insights into the underlying neurophysiological mechanism when using these concept generation techniques.
Game-based learning has been shown to motivate and engage students in the domain of learning, knowledge and thinking skills, such as problem-solving and decision-making. This article reports the design and use of an educational board game for learning chemistry that attempts to improve students’ scientific concepts as well as creative problem solving (CPS) skills. The investigation of a board game as a teaching material was conducted in a field test with 48 high school students. After experiencing gameplay, most students’ CPS skills increased, especially in the construct of solution-finding. Their scientific concepts of chemical techniques and products also improved significantly according to the comparisons of pre- and post-test results, indicating that the game context can help students develop a holistic view of the function of chemistry knowledge. Student interviews revealed the nuances of the improvements in the conceptual dimensions as well as the interactivity to construct creative ideas from different points of view. The recommendation to integrate science-related social issues into the game mechanism to enable students to explicitly experience all stages of the CPS process in each game round is the main outcome of this research.