A theoretical user experiences design model and allied implementation with AIGC and autonomous agent growing scheme
Xiufeng Lin, Wei Jin, Qi Xin, Hongzhi Chen, Shijia Gu · 2025
By responding to the complexity of user experience (UX) design among the enterprise information (2B) systems, such as role and character distinct, organization isolated, simplified but uniform human-machine interaction logic etc, the manuscript introduces a theoretical model for dealing with optimal UX design among complex 2B systems. Since the classical UX design were strongly rely on designers’ experience and understanding over the UX design optimal, an autonomous agent framework which can be allied implemented by using the generative large-scale AI model is engaged, with an original design of the growing scheme, the superior agent individual can be extracted from the competition and voting over the agent cluster, which can be considered as the agent individual that best understood the optimal UX design model. In order to dig deeper regarding to the potential of advance individual, an attention spreading scheme is invented and applied to the advanced agent to trigger more innovation abilities. By applying the product design data including screenshot, product documentations, and user logging from various 2B systems as the reinforced learning of growth agents, the evaluating results demonstrate the reasonable of UX design optimal model proposed, feasibility of autonomous agent growing scheme, as well as usefulness of original designed attention spreading mechanism for UX design among complex 2B system, especially with sparse data sample.