gen-briefly-explain-self-information-and-one-common-misconceptio-be03

Briefly explain self-information and one common misconception about it.

## Self-Information **Self-information** (or surprisal) measures how "surprising" or "informative" a specific event is. It's defined as: I(x) = -log₂ P(x) where P(x) is the probability of event x. The result is measured in bits. **Key idea:** Rare events (low probability) have high self-information, while common events (high probability) have low self-information. ## Common Misconception **Misconception:** Self-information measures the "importance" or "usefulness" of information. **Reality:** Self-information is purely a function of probability—it measures statistical surprise, not semantic value or practical usefulness. A highly improbable but meaningless event (like a specific random noise pattern) has high self-information, while a rare but crucial message might have the same self-information as any other equally probable event. The measure is objective and mathematical, independent of context or human judgment about what matters.