Model Type: Supervised Classification Model
What it does: Recognizes and sorts things into categories based on example data.
Media: Sound and Images

A classifier is one of the simplest types of AI models. It learns to recognize patterns in data and sort them into categories you define.

It can answer questions like:
“Is this a cat or a dog?”
“Is this sound a laugh or a sneeze?”
“Does this image belong to style A or style B?”

You train it by giving it labeled examples — the more the better — and it learns what makes each category unique.

Workings

1. Pick a classification task

Think of 2–5 categories you want to distinguish. For example:

• “Happy / Sad / Confused” facial expressions
• “Wobbly line / Geometric line / Chaotic line”
• “Voice: whisper / normal / shout”
• “Plant leaf shapes: round / jagged / split”

2. Collect data for each category

• At least 20–30 samples per category, ideally more
• Use your phone, camera, drawing pad, or microphone
• Make sure labels are consistent and clear

3. Train the model

• Use Teachable Machine (easy and browser-based)
• Choose Image, Audio, or Pose classifier
• Upload your labeled data
• Press “Train Model”

4. Test and use your model

• Try real-time testing via webcam, mic, or file upload
• Export the model to use in websites or installations
• Reflect on what works and what fails — and why

Why Try a Classifier?

• Fastest way to train a working model
• No coding required (with Teachable Machine)
• Great for testing the boundaries of machine understanding
• Works with any kind of media: drawings, sounds, gestures, objects
• Excellent for designing your own categories and exploring bias

Resources

Teachable Machine