log loss

Log loss (also called binary cross-entropy or logarithmic loss) is a loss function that measures the performance of a probabilistic binary classifier. Given true labels y∈{0,1} and predicted probabilities p∈(0,1), log loss = -[y·log(p) + (1-y)·log(1-p)]. It is minimized when predictions exactly match true labels and increases sharply as predictions diverge, especially for high-confidence errors. The function operates on the open interval (0,1) — predictions at exactly 0 or 1 produce infinite loss, which enforces probabilistic calibration. Persistence mechanism: encoded in optimization libraries and evaluation frameworks (scikit-learn, TensorFlow, PyTorch) as a standard metric for binary classification tasks. [formal: log loss | substrate: mind | horizon: a moment | explicit: yes | epoch: 0.92]