Teaching Reform of Machine Learning Courses Based on a Multi-stage Cognitive Advancement Framework

In recent years, machine learning courses have been widely introduced in undergraduate and graduate programs such as Artificial Intelligence, Computer Science and Technology, Data Science and Big Data Technology at major universities. Due to the inherent complexity of machine learning, both students and teachers face high demands in the teaching and learning process. Moreover, as a course established under the context of New Engineering, it lacks mature and exemplary experiences. This results in many contradictions in the current teaching process of machine learning, the contradiction between the singularity of teachers’ knowledge structures and the multidisciplinary intersectionality of machine learning; the contradiction between the limitations of students’ knowledge reserves, the simplified and superficial description of teaching resources and the multi-level nature of machine learning knowledge. To address these issues, this paper proposes a multi-stage cognitive advancement framework for curriculum reform. It manifests as vertical advancement in knowledge within the classroom, horizontal advancement between different areas of knowledge, and advancement from classroom to practice. It covers the entire teaching process—from theory to practice, from in-class to after-class, and from offline to online—aiming to enhance teaching effectiveness and better cultivate students’ thinking and practical abilities. Experimental evidence shows that the curriculum reform of machine learning based on the multi-stage cognitive advancement framework can effectively alleviate the existing contradictions, comprehensively improve teaching effectiveness, and cultivate talents in the field of artificial intelligence that meet the requirements of the new era.