Id: arxiv/2611.18511 Url: https://arxiv.org/abs/2611.18511 Tags: cs.LG, math.NA, stat.ML Score: 0 Title: On numerical issues of divergences between nearly equal distributions Domain: cs.LG Source: arxiv Authors: Brandon Amos Reviews: 0 Subject: cs.LG Category: research Comments: 0 Disputed: false Division: cs Mentions: 0 Discussed: 0 Published: 2026-11-13 Page: https://delphi.science/r/arxiv/2611.18511 Barron et al. (2026) recently identified numerical issues in the softmax and soft cross-entropy loss. We show that similar issues are present in standard divergence operations between categorical distributions. This includes the forward and reverse Kullback-Leibler (KL) divergence and generalized Jensen-Shannon divergence (JSD), as well as Monte Carlo KL estimates and entropy. When the distributions are nearly equal, rounding and cancellation issues affect the forward and gradient computations. These issues can affect settings using divergences in optimization objectives, such as knowledge…