DesignMind ML-personalized UI/UX design with adaptive color and simplified features

Read full text Summarise with VIRAT
Authors: Rakhi Rani Paul, Sunjida Akter Shanu, Maria Islam, Nahian Fairuz, Subrata Kumer Paul, Mahfuzur Rahman, Md. Ekramul Hamid
Cited by: 0
References: 22

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Personalized user interface and user experience (UI/UX) design has become increasingly important for enhancing usability, accessibility, and user satisfaction in modern digital systems. Traditional static interfaces often struggle to accommodate diverse user preferences and interaction behaviors, underscoring the need for intelligent adaptive interface systems. This study aims to develop a machine learning–based framework that dynamically personalizes UI/UX design by analyzing user interaction behavior, visual interface characteristics, and the evolution of temporal preferences. To achieve this objective, we propose DesignMind-ML, a multimodal framework that integrates Artificial Neural Networks (ANN), Convolutional Neural Networks (CNN), and GRU-LSTM models. The ANN learns instruction-level UI adaptation decisions, such as color preference and feature prioritization; the CNN evaluates UI screenshots to assess visual layout and complexity; and the GRU-LSTM captures sequential user behavior patterns over time. The outputs of these models are combined using a weighted decision-level integration strategy to generate adaptive interface configurations. The system was evaluated using a self-constructed multimodal dataset of 3,000 user interaction sessions. Experimental results indicate that the hybrid framework achieves approximately 76% overall UI adaptation accuracy, demonstrating the feasibility of multimodal machine learning for intelligent and behavior-aware UI personalization in adaptive digital systems.

Fields of Study & Subjects

Usability and User Interface Design Recommender Systems and Techniques Innovative Human-Technology Interaction

Publication details

Journal / source
Computer Science and Engineering Research
Volume
3
Issue
1
Pages
19–27
Publication date
27 March 2026
Language
EN