Thesis

Modelling learning in design

Creator
Rights statement
Awarding institution
  • University of Strathclyde
Date of award
  • 2000
Thesis identifier
  • T10055
Department, School or Faculty
Abstract
  • Learning in design is a phenomenon that exists in the domain of design. That the phenomenon exists is evident from the many systems built based on Artificial Intelligence and machine learning techniques in an attempt to support designers at work. However, the development of these systems may not be well founded on a holistic basis as evident from the literature review. Hence, this thesis contributes to the knowledge of modelling learning in design and has resulted in a novel and comprehensive model.To model learning in design, formalisms for design and learning activities were derived based on the corpus of research work published. To explain the inextricable link between learning and design, three hypotheses were raised. These hypotheses posited that the cognitive activities of design and learning are linked at the epistemic level, teleological level and temporally. The formalisms and hypotheses were substantiated through a protocol analysis of a recording of a designer at work. The model based on the formalisms and hypotheses was presented for peer critique at one of the Machine Learning in Design (MLinD) workshops. Based on the outcome the protocol analysis and the peer review, the model was further evolved, resulting in the model called Model-LinD.Model-LinD features learning in design as the transformation of input knowledge into output knowledge that is activated by rationale trigger(s) and through the interactions between the design and learning activities that are determined by the epistemic, teleological and temporal links. The evolved model was evaluated against a set of evaluation criteria, which showed that it is indeed a model capable of explaining the phenomenon of learning in design. Hence, this is claimed as the main contribution of this thesisIn order to derive a theory of learning in design from Model-LinD, the strengths and weaknesses of deriving Model-LinD have been discussed leading to some suggestions for further work.
Advisor / supervisor
  • Duffy, Alex H. B. (Alex Hynd Black), 1957-
  • Gay, Robert
Resource Type
EThOS ID
  • uk.bl.ethos.248557

Relations

Items