Little Bits of Artificial Intelligence

Step 1: Counting and Comparing

Goal: Learn to count how many items (like words) match.

• What to Practice:
• Count items in two groups.
• Compare if two words/sentences share the same items.
• Real-World Example:

“I have 3 apples, you have 2 apples. How many do we both have in common?”

Step 2: Understanding Sets (Groups of Words)

Goal: Learn intersection and union of word sets.

• Key Terms:
• Intersection: What’s common in both?
• Union: All unique items from both.
• Practice:

Word set A = {hi, there}
Word set B = {there, you}

• Intersection = {there}
• Union = {hi, there, you}

Step 3: Simple Division and Ratios

Goal: Learn how to compare parts vs whole.

• Concept:

If 2 words match out of 5 total, that’s a confidence score:
2 ÷ 5 = 0.4 (40% confident)

• Practice:
• Fractions (e.g., 3/4, 2/5)
• Turn fractions into percentages (e.g., 0.4 → 40%)

Step 4: Text Cleaning and Matching

Goal: Learn how to simplify words (like “running” → “run”).

• What to Learn:
• Remove endings like ing, ed, es, s
• Treat similar words as same root word (basic stemming)
• Why It Matters: Helps match more words across sentences.

Step 5: Confidence Scoring

Goal: Use previous math to decide how sure we are.

• How to Calculate:
• Confidence = Common Words / Total Unique Words
• Higher score = better match!

Step 6: Label Guessing (Semi-Supervised Learning)

Goal: Learn how the model guesses labels from a few examples.

• How It Works:
• Compare new (unlabeled) sentence to known (labeled) ones.
• Use similarity (confidence) to pick the best matching label.
• Add that guessed label to the data.
• Example Flow:
• Known: “Hi” → Greeting
• Unknown: “Hello there” → similar → Guess = Greeting

Learning Path Summary Table