Research on the Design of Virtual Simulation Teaching System for Ideological and Political Courses under the Framework of Deep Learning

Aiming at the problems of insufficient innovation in ideological and political teaching and low student participation, this study designs and implements a virtual simulation teaching system for ideological and political courses based on a deep learning framework, and proposes a teaching guidance algorithm that integrates reinforcement learning and emotional computing. The system adopts a layered architecture, covering core modules such as virtual scenes and teaching guidance. The algorithm perceives students' learning status and emotions in real time through emotional computing, and dynamically adjusts teaching strategies in combination with reinforcement learning. The experiment selected 200 students and divided them into an experimental group and a control group for an 8-week teaching experiment. The results showed that the average score of the ideological and political knowledge test in the experimental group reached 88.6 points, which was significantly higher than the 76.3 points of the control group, with an increase of 16.1%; the average number of interactions in the experimental group was 13.2 times per week, which was 2.28 times that of the 5.8 times in the control group; in terms of emotional feedback, the proportion of positive emotions in the experimental group increased from 40% before the experiment to 72%, while that in the control group only increased from 38% to 43%. Experiments show that the system and algorithm effectively improve the effect of ideological and political teaching, student participation and emotional investment, and provide a new path for ideological and political teaching reform.