Dimensions of machine learning in design

This paper discusses the effects of direct and indirect communications on social learning and task coordination in design teams. The findings reported in this paper are based on a computational model that simulates the formation of transactive memory (TM) through social learning from direct and indirect communications. Direct communications are explicit information exchanged between team members whereas indirect communication may be opportunistic and coincidental, resulting in learning and information gained through observations of the actions of others. However, team structure mediates opportunities for communication. Three types of team structures are studied, which are differentiated on the basis of their constraints on and opportunities for direct and indirect communications across the team. The differences across the team structures are investigated through a series of simulations in which team member retention, cognitive busyness of team members, and task complexity are additional moderating variables, and task coordination and formation of TM are the dependent variables. Fewer communications to coordinate the same tasks are taken as the measure of efficient task coordination. Findings suggest that reduction in communication and learning opportunities are more detrimental to the task coordination in flat teams as compared to functional teams. Indirect communications contribute more to the formation of TM than to task coordination. Flat teams facilitate the formation of TM, whereas functional teams are more appropriate for efficient task coordination, indicating that the role of TM in mediating task coordination varies with team structure.

Section: Introductionmentioning

confidence: 99%

Social learning in design teams: The importance of direct and indirect communications

Singh

Dong

Gero

2013

“…This section attempts to categorize our work on using machine learning to initialize an optimization in terms of the "dimensions" of machine learning in design presented in Grecu and Brown (1996).…”

Section: Dimensions Of Machine Learning In Designmentioning

confidence: 99%

“…The decision on when to run the learner was thus one of human judgment. Learners that are invoked manually should be added to Grecu and Brown's (1996) taxonomy. We envision a system that uses our techniques to learn incrementally (see Section 8).…”

Section: What Can Trigger Learning?mentioning

confidence: 99%

Learning to set up numerical optimizations of engineering designs

Schwabacher

Ellman

Hirsh

1998

Gradient-based numerical optimization of complex engineering designs offers the promise of rapidly producing better designs. However, such methods generally assume that the objective function and constraint functions are continuous, smooth, and defined everywhere. Unfortunately, realistic simulators tend to violate these assumptions, making optimization unreliable. Several decisions that need to be made in setting up an optimization, such as the choice of a starting prototype and the choice of a formulation of the search space, can make a difference in the reliability of the optimization. Machine learning can improve gradient-based methods by making these choices based on the results of previous optimizations. This paper demonstrates this idea by using machine learning for four parts of the optimization setup problem: selecting a starting prototype from a database of prototypes, synthesizing a new starting prototype, predicting which design goals are achievable, and selecting a formulation of the search space. We use standard tree-induction algorithms (C4.5 and CART). We present results in two realistic engineering domains: racing yachts and supersonic aircraft. Our experimental results show that using inductive learning to make setup decisions improves both the speed and the reliability of design optimization.

“…In this section, a classification of the work according to a proposed scheme and a critique of the scheme is presented (Grecu & Brown, 1996).…”

Section: Categorization Of the Research And Critique Of Proposed Catementioning

confidence: 99%

Machine learning for simulation-based support of early collaborative design

Ivezic

Garrett

1998

The research and development of a simulation-based decision support system (SB-DSS) capable of assisting early collaborative design processes is presented. The requirements for such a system are included. Existing collaborative DSSs are shown to lack the capability to manipulate complex simulation-based relationships. On the other hand, advances within the machine learning in design community are shown to have a potential for providing, but have not yet addressed, simulation-based support for collaborative design processes. The developed SB-DSS is described in terms of its four principal components. First, the behavior-evaluation (BE) model is used to both structure individual, domain-specific decision models and organize these models into a collaborative decision model. Second, a probabilistic framework for the BE model enables management of the uncertainty inherent in learning and using simulation-based knowledge. Significantly, this framework provides a constraint satisfaction environment in which simulation-based knowledge is used. Third, a statistical neural network approach is used to capture simulation-based knowledge and build the probabilistic behavior models based on this knowledge. Fourth, since a probability distribution theory does not exist for the nonlinear neural network approaches, Monte Carlo simulation is introduced as a method to sample the trained neural networks and approximate the likelihoods of design variable values. Consequently, constraint satisfaction problem-solving capability is obtained. In addition, a mapping of the SB-DSS architecture onto a collaborative design agent framework is provided. Experimental evaluation of a prototype SB-DSS system is summarized, and performance of the SB-DSS with respect to search and usability metrics is documented. Initial results in developing the simulation-based support for collaborative design are encouraging. Lastly, a categorization of the machine learning approach and a critique of the proposed categorization scheme is presented.