Learning a Metric Embedding for Face Recognition using the Multibatch Method
Oren Tadmor, Tal Rosenwein, Shai Shalev-Shwartz, Yonatan Wexler, Amnon Shashua
–Neural Information Processing Systems
We first show that the estimate is unbiased. Next, we turn to analyze the variance of the multibatch estimate. Let us momentarily fix r and omit the superscript from A(r). For q =1: we obtain E A2 (i), (j) which is the variance of the random variable rr`i,j(z) rrL(z). The expression we obtain is E A (i), (j)A (i), (t).
Neural Information Processing Systems
Apr-22-2026, 08:42:48 GMT