Interview, Fireside Chat
Judea Pearl: Correlation and Causation | AI Podcast Clips
- Researchers plan to detect correlations in observational studies where drivers choose autonomous features to infer causal relationships regarding system-induced sleepiness in semi-autonomous vehicles.
- Scientists anticipate imposing causal logic on correlations within psychology and disciplines with hard-to-account-for variables, a practice predicted to become obsolete if current methodologies persist.
- Applied psychology regarding human behavior in semi-autonomous vehicles is expected to become outdated if it continues relying on inferences from correlation to causation.
- Current scientific understanding was orphaned for centuries until the 1920s, when mathematics capable of capturing the specific asymmetry of causation was finally developed.
- Future scientific advancement requires developing mathematical tools to capture directional causation, as existing physics equations remain symmetrical and cannot distinguish between X causing Y and Y causing X.