Retracted: Self-Evolving Neural Intelligence Tutor for Competency-Based Skill Development Using Curriculum-Aware Adaptive Learning Mechanisms

Intelligent Tutoring Systems (ITS) have been developed significantly by the rapid evolution of AI in the education sector, although most of the existing ones do not offer dynamic adaptability, curriculum integration, and scalable personalization. To overcome these, this paper proposes a Self-Evolving Neural Intelligence Tutor which integrates both Deep Q-Networks (DQN) and Curriculum-Aware Knowledge Graphs in terms of competency-based skill development. The system has been deployed in Python based on the TensorFlow framework and analyzes the interactions of the learners, provides an initial profile of their proficiency and does dynamic recommendations on the learning materials on Beginner, Intermediate and Advanced levels. On the publicly available English Learning ITS Performance Dataset on Kaggle, the proposed model yielded 93.2 % classification accuracy and a lower RMSE of 0.215, which is an average 7.8 higher than the results of modern adaptive learning methods. These findings exhibit improved performance, high reliability and significant error reduction. The model promotes evolving continuously and adapting on a real-time basis, which makes it applicable in the context of scalable learner-centered digital education platforms.