[初賽第二階段] 改進人工智慧系統（2） [Preliminary Phase 2] Improving AI Systems (2)

Big Concept: Improving AI Systems Topic: RAG vs. Fine-Tuning Bloom's Taxonomy: remember, understand, apply, analyze Learning Goal: Deliver a clear lecture explaining two common ways to improve large language model applications: Retrieval-Augmented Generation and fine-tuning. First, help students understand the basic difference between the two methods: RAG gives the model access to external information at answer time, while fine-tuning changes the model’s behavior through additional training. Then guide students to apply this distinction to real-world scenarios. Explain when RAG is more suitable and when fine-tuning is more suitable. Use real-world examples to illustrate these choices. Key Learning Points: RAG; model parameter fine-tuning; when to use RAG; when to use fine-tuning