This session illustrates the fundamental difference between classical (rule-based) programming and machine learning using simple diagrams and real-life examples.
In traditional programming:
- The developer provides rules (logic) and data.
- The computer produces answers based on the given logic.
📌 Example:
A spam filter uses explicit rules like:
if "lottery" or "free money" in email: mark as spam
This works but is rigid—slight variations bypass the rules.
In machine learning:
- You provide data and answers (labels).
- The algorithm learns rules (model) from them.
📌 Example: A spam classifier learns patterns from thousands of labeled spam and non-spam emails. Over time, it generalizes and detects spam—even without predefined rules.
A face recognition system doesn’t get programmed with features like “nose” or “eye.” Instead, it learns features by analyzing thousands of images and recognizing pixel patterns.