The public visibility of artificial intelligence (AI) is rising rapidly, driven by its positive impact across many disciplines. Consequently, courses that cover AI and machine‑learning fundamentals are essential for grasping their societal role and potential. Effective teaching must tightly integrate theory with practice to master core concepts and basic algorithms. This essay shares our experience designing machine‑learning exercises for automated programming assessment platforms. We are not creating a new grading system; instead, we propose viewing machine‑learning problems as input‑output assessment tasks. Under this view each exercise has a unique, deterministic answer, allowing tools such as VPL for Moodle, Codeforces, and MOJ to support AI education efficiently. The approach encourages instructors to adopt more dynamic, interactive teaching methods that deeply blend theory and practice, improving learning outcomes for AI, especially machine learning.
Review