Model prediction uncertainty
Uncertainty in model prediction plays an important role in affecting decision-making activities, and the quantified uncertainty is closely associated with risk assessment. In particular, uncertainty in model prediction underpins many crucial decisions related to life or safety- critical applications [73].
ENTITY
2 - AI
INTENT
2 - Unintentional
TIMING
3 - Other
Risk ID
mit339
Domain lineage
7. AI System Safety, Failures, & Limitations
7.3 > Lack of capability or robustness
Mitigation strategy
1. Implement rigorous Uncertainty Quantification UQ methods to estimate and explicitly report prediction intervals e.g., confidence or credible intervals alongside point estimates to inform critical evaluation and decision-making reliability. 2. Employ advanced techniques, such as Bayesian Neural Networks or Monte Carlo Dropout, to enhance model robustness, mitigate epistemic uncertainty stemming from model choice, and improve prediction stability. 3. Establish clear protocols for human-in-the-loop oversight, triggering human review or intervention when the quantified uncertainty of a prediction exceeds a predefined, risk-appetite-aligned threshold, particularly in safety-critical applications.