From Black Box to Explainability: An AIIntegrated and Value-Oriented Pedagogical Reform for Undergraduate Applied Stochastic Processes in Statistics Programs

Abstract

The rapid advancement of artificial intelligence is reshaping higher education, while value-oriented education (curriculum-based ideological and political education) has become a fundamental requirement for talent cultivation in the new era. Applied Stochastic Processes, a core course for statistics majors, is characterized by abstract theory, intensive mathematical derivations, and broad applications. Traditional teaching methods face three structural dilemmas: heavy emphasis on mathematical derivations with little connection to real-world applications, focus on computational techniques at the expense of conceptual understanding, and rigid model assumptions without fostering critical thinking. This paper proposes a dual-driven pedagogical reform framework integrating AI and value-oriented education, using “explainability” as the cognitive bridge that organically connects stochastic process theory with AI methods while embedding value guidance. Utilize AI visualization tools to lower the cognitive threshold of stochastic processes, introduce large language model-assisted programming accompanied by a ‘critical programming’ strategy to prevent over-reliance; integrate the spirit of scientists into teaching to enhance cultural confidence; and use Markov chains and the PageRank algorithm as a case study to embed data ethics and social responsibility education into the explanation of algorithm principles. Teaching practice demonstrates that this reform effectively enhances students’ conceptual understanding, programming skills, and dialectical thinking abilities, achieving an organic unity of knowledge transmission, competency development, and value cultivation.