How BrainTube Automatically Extracts Key Concepts

Every video mentions dozens of concepts. BrainTube identifies them, defines them, and tracks them across your library.

What Are Concepts?

Concepts are the building blocks of knowledge:

• "Compound Interest" (finance)
• "Cognitive Load" (psychology)
• "Technical Debt" (engineering)

When BrainTube processes a video, it extracts:

• Term: The concept name
• Definition: What it means in context
• First mention: Timestamp where it appears
• Frequency: How many times it's mentioned

Why This Matters

1. Build Vocabulary Automatically

As you save videos, your concept library grows:

• 10 finance videos → 50+ finance terms defined
• 10 ML videos → 40+ ML terms defined

Each term links back to the video where you learned it.

2. See Concepts Across Videos

The same concept appears in different contexts:

"Feedback loop" mentioned in:

• Systems thinking video (at 4:23)
• Product management video (at 12:45)
• Habit formation video (at 8:30)

You see how one concept applies everywhere.

3. Track Your Learning Depth

Surface learning: 5 concepts from one video

Deep learning: 50 concepts across 10 videos, interconnected

The Knowledge Graph shows which concepts connect to others.

How to Use Concepts

Browsing Concepts

Open any video → Click "Concepts" tab → See all extracted terms with definitions

Searching by Concept

Search for any term. BrainTube finds:

• Every video mentioning it
• The timestamp of first mention
• Your highlights related to it

Connecting to Interests

Concepts auto-link to your Interest areas:

• "Gradient Descent" → linked to Machine Learning interest
• "P/E Ratio" → linked to Finance interest

Real Example

Video saved: "Introduction to Machine Learning" (30 min lecture)

Concepts extracted:

• Supervised Learning — Training models with labeled data
• Overfitting — Model performs well on training data but poorly on new data
• Feature Engineering — Selecting and transforming input variables
• Cross-Validation — Technique for evaluating model performance
• Gradient Descent — Optimization algorithm for minimizing loss

Each concept has a timestamp so you can revisit the explanation.

Pro Tip

When you encounter a concept you already know, notice how different experts define it. This deepens your understanding and reveals nuances you might have missed.