From Traditional GUI to Intelligent UI: Building User‑Centred Experience

This article explores how graphical user interfaces (GUI) evolve from traditional, static designs toward intelligent, user‑experience‑driven interfaces.

It introduces a three‑layer model for UI design: the first layer inherits precision, usability and performance from classic software engineering; the second layer adds intelligent UI capabilities that influence user behavior; the third layer leverages emerging technologies such as AR/VR and machine learning to create novel interaction experiences.

Using Shannon’s information theory, the author quantifies the information load of UI elements, showing how excessive or insufficient visual data can increase user effort and reduce conversion. The analysis demonstrates how to measure and optimize the effective coding length of UI components, especially in complex e‑commerce pages like Taobao’s “Juhuasuan” channel.

The paper then discusses practical UI control techniques, including A/B experiments that adjust visual layout (V) and interaction flow (I) under constraints derived from product‑manager insights. By applying the maximum‑entropy principle, the author derives probability distributions for user clicks and shows how to align them with business goals.

Finally, the article looks ahead to future trends: deeper integration of AI assistants, multimodal input (gesture, voice, pose), and the convergence of digital and physical worlds through standardized service APIs. It argues that the ultimate direction of UI intelligence is a shift from visual widgets to binary, service‑oriented interactions powered by personal AI agents.