Results for "positive predictive value"

Of predicted positives, the fraction that are truly positive; sensitive to false positives.

Often more informative than ROC on imbalanced datasets; focuses on positive class performance.

Of true positives, the fraction correctly identified; sensitive to false negatives.

Plots true positive rate vs false positive rate across thresholds; summarizes separability.

Expected cumulative reward from a state or state-action pair.

AI predicting crime patterns (highly controversial).

Fundamental recursive relationship defining optimal value functions.

Scalar summary of ROC; measures ranking ability, not calibration.

A table summarizing classification outcomes, foundational for metrics like precision, recall, specificity.

Activation max(0, x); improves gradient flow and training speed in deep nets.

Optimal control for linear systems with quadratic cost.

Stores past attention states to speed up autoregressive decoding.

Combines value estimation (critic) with policy learning (actor).

Maximum expected loss under normal conditions.

Learning by minimizing prediction error.

Average of squared residuals; common regression objective.

Normalized covariance.

Returns above benchmark.

Expected return of taking action in a state.

Sample mean converges to expected value.

Optimizing policies directly via gradient ascent on expected reward.

Approximating expectations via random sampling.

Model optimizes objectives misaligned with human values.

Directly optimizing control policies.

Inferring and aligning with human preferences.

Optimizes future actions using a model of dynamics.

High-fidelity virtual model of a physical system.

Systematic error introduced by simplifying assumptions in a learning algorithm.

Penalizes confident wrong predictions heavily; standard for classification and language modeling.

A point where gradient is zero but is neither a max nor min; common in deep nets.