Results for "learning signal"
Failure to detect present disease.
Grouping patients by predicted outcomes.
Predicting disease progression or survival.
Differences between training and deployed patient populations.
Unequal performance across demographic groups.
AI predicting crime patterns (highly controversial).
Models estimating recidivism risk.
Predicting case success probabilities.
Ultra-low-latency algorithmic trading.
Predicting borrower default risk.
Ensuring models comply with lending fairness laws.
AI discovering new compounds/materials.
AI proposing scientific hypotheses.
Fast approximation of costly simulations.
Agents have opposing objectives.
System-level design for general intelligence.
Tendency to gain control/resources.
Internal representation of the agent itself.
Restricting distribution of powerful models.