Negative Log Likelihood

[!def] Negative Log Likelihood (NLL)
$$
NLL = - log (\prod_i \hat{y_i} ^ {y_i})
$$
$$
NLL = - (\sum_i y_i ; log : \hat{y_i})
$$

• Negative log likelihood is the negative-log of the Likelihood
• Also, known as Multi Class Cross Entropy