LOTI: Using Machine Learning to simulate subjective opinions in design

Abstract

The objective of the workshop described in the article was to redesign a chair called Loti. In a subjective opinion shared by the authors and the participants of the workshop, the chair seems plagiarism of a famous chair by Ray and Charles Eames. The authors centralised the workshop on the use of computational tools for assessing subjective opinions. The authors and the participants created a method for detecting plagiarism and implemented it in the process of design. They created a parametric model of the chair that allowed changing the chair's components with variables. Using this model, the participants generated multiple variations and surveyed other students to assess which of the versions seemed plagiarism. With the information obtained from the survey, we trained a neural network to relate the variables with the level of plagiarism. We linked the parametric model with the neural network to create a tool that informs the user about the probability of committing plagiarism in real-time. The participants used the tool for designing new chairs to evaluate the efficiency of the method.