Development Of Curriculum Ideology And Politics Resource Recommendation And Management Platform Based On Deep Reinforcement Learning

Curriculum ideology and politics refer to the integration of ideological and political content into the teaching of other courses,enabling non-ideological and political courses to also serve as vehicles for ideological and political education.Reinforcement learning is an interactive learning method.When compared with static recommendation models,the interactive update feature of reinforcement learning better addresses the issue of user interest shift in the field of recommendations.Addressing the dearth of ideological and political cases in the current curriculum’s ideological and political teaching,this paper optimizes the deep reinforcement learning recommendation algorithm.It develops a curriculum ideological and political case resource management platform and applies the optimized algorithm to enable dynamic recommendation of curriculum ideological and political cases.The main contributions of this work are as follows:(1)The interaction between users and the recommendation system is modeled as a Markov decision-making process.Building upon the classical reinforcement learning algorithm Actor-Critic,we propose a deep reinforcement learning recommendation algorithm optimized for recommendation diversity.We achieve this by optimizing the loss function of the algorithm through the incorporation of model entropy,thereby enhancing the diversity of recommendation results.(2)Building upon the aforementioned algorithm,we design a reward prediction model to optimize the agent’s interactive update process during the strategy update.By doing so,we propose a deep reinforcement learning recommendation algorithm optimized for sample efficiency.This leads to accelerated training speed of the model and improved sample efficiency in reinforcement learning.(3)Based on an analysis of the current state of curriculum ideological and political resources and the needs of teachers and users,we design and develop a curriculum ideological and political resource management platform using technologies such as containers,caches,and reverse proxy.This platform facilitates the retrieval and browsing of curriculum ideological and political resources.(4)Leveraging the optimization algorithm proposed in this paper,we develop a recommendation system based on deep reinforcement learning for the curriculum ideological and political resource dataset within the platform.By integrating the recommendation system with the platform,we provide teachers with course ideological and political case recommendations.Finally,we test the functionality and performance of the platform,and the results demonstrate its ability to provide an excellent user experience and high availability.