Direct Loss Minimization (DLM) has been proposed as a pseudo-Bayesian method motivated as regularized loss minimization. Compared to variational inference, it replaces the loss term in the evidence lower bound (ELBO) with the predictive log loss, which is the same loss function used in evaluation. A number of theoretical and empirical results in prior work suggest that DLM can significantly improve over ELBO optimization for some models. However, as we point out in this paper, this is not the case for Bayesian neural networks (BNNs). The paper explores the practical performance of DLM for BNN, the reasons for its failure and its relationship to optimizing the ELBO, uncovering some interesting facts about both algorithms.

翻译：直接损失最小化(DLM)建议作为一种假贝耶斯方法,其动机是将正常损失最小化。与变式推论相比,它将证据下限(ELBO)中的损失术语替换为预测日志损失,这与评价中使用的损失功能相同。先前工作的一些理论和经验结果表明,DLM可以大大改进某些模型的ELBO优化。然而,正如我们在本文件中指出的那样,Bayesian神经网络(BNNS)的情况并非如此。该文件探讨了BNN(BN)DLM的实际性能、失败的原因及其与优化ELBO的关系,揭示了关于这两种算法的一些有趣的事实。