User Interface Optimization using Genetic Programming with an Application to Landing Pages

Abstract

The design of user interfaces (UIs), such as World Wide Web pages, usually consists in a human designer mapping one particular problem (e.g., the demands of a customer) to one particular solution (i.e., the UI). In this article, a technology based on Genetic Programming is proposed to automate critical parts of the design process. In this approach, designers are supposed to define basic content elements and ways to combine them, which are then automatically composed and tested with real users by a genetic algorithm in order to find optimized compositions. Such a strategy enables the exploration of large design state-spaces in a systematic manner, hence going beyond traditional A/B testing approaches. In relation to similar techniques also based on genetic algorithms, this system has the advantage of being more general, providing the basis of an overall programmatic UI design workflow, and of calculating the fitness of solutions incrementally. To illustrate and evaluate the approach, an experiment based on the optimization of landing pages is provided. The empirical result obtained, though preliminary, is statistically significant and corroborates the hypothesis that the technique works.

Crowdsourcing Interface Feature Design with Bayesian Optimization

“…Clearly this encompasses the much broader workflow of making design changes based on feedback and controlled experiments [9]. More relevant to the context of this paper, however, are efforts that formalize this approach [13,23,28]. Salem [23] demonstrates a structured approach to comparing and refining web landing page design alternatives using genetic programming while Liu et al [13] explore optimal representations for mathematics pedagogy.…”

EUDoptimizer: Assisting End Users in Composing IF-THEN Rules Through Optimization

“…Applying optimization methods to user interface design is a long standing topic in human-computer interaction research: when assumptions are appropriate, optimization methods offer a greater-than-zero chance of finding an optimal design [9]. Optimization methods have been firstly adopted for keyboard layouts [10], [11], and then in many other application areas, e.g., accessibility [12], menus [13], [14], sketching [15], and web layouts [16]. Since EUD interfaces are typically organized through grid menus, our work is strictly related to the menu optimization problem.…”

Introduction to Intelligent User Interfaces (IUIs)

“…This research can be used to help designers create intelligent user interfaces for HCI decision-making applications that respond to user experience and decision-making efficiency. In [19], a Genetic Programming-based technology is proposed for automating crucial design phases. Designers can specify simple content elements and ways to merge them in this method, which will then be automatically composed and checked with actual users by a genetic algorithm to find optimal compositions.…”