Adaptive User Interface for Smart Programming Exercise

Yutaka Watanobe, Md. Mostafizer Rahman, Alexander Vazhenin, Jun Suzuki

Abstract: An adaptive user interface for smart programming exercise and its platform is presented. The proposed adaptive user interface is oriented to repetitive exercises with many pro-gramming tasks through different learning phases. The learning phases include searching, reading, coding, testing, debugging, and refactoring, and the learner can receive smart assistance in each phase. The content of the smart assistance can be adjusted by modes predefined for each phase. The configuration and contents of the user interface are controlled by the system according to the transition of the learner's state and activities. The smart assistance is realized by different types of materials and automatic assessment systems for program codes as well as by machine learning models for the recommendation, code completion, bug highlighting, program repairing, and program transformation. In this paper, the state transition graph to organize the adaptive user interface and its smart assistant modes for each learning phase are presented. The prototype of the user interface as well as the architecture of its platform including the automatic assessment system, different machine learning models, and Iogging ecosystem are also demonstrated.