arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2007.03293cs.LGstat.ML
收藏

Single Shot MC脱落近似

Single Shot MC Dropout Approximation

Elektrobit Automotive GmbH;Zurich University of Applied Sciences(苏黎世应用科学大学);University of Zurich(苏黎世大学);Konstanz University of Applied Sciences(康斯坦茨应用科学大学)

Kai Brach, Beate Sick, Oliver Dürr

摘要:

深度神经网络(DNN)以其高预测性能而闻名,特别是在物体识别或自动驾驶等感知任务中。尽管如此,DNN在遇到全新的情况时,在没有表明其不确定性的情况下,很容易产生不可靠的预测。DNN(BDNN)的贝叶斯变体,如MC dropout BDNN,确实提供了不确定性度量。然而,BDNN在测试期间很慢,因为它们依赖于采样方法。在这里,我们提出了一个Single Shot MC dropout近似,它保留了BDNN的优点,而不会比DNN慢。我们的方法是解析近似的每一层在一个完全连接的网络的期望值和方差的MC dropout信号。我们评估我们的方法在不同的基准数据集和模拟玩具的例子。我们证明了我们的Single Shot MC dropout近似类似于用MC方法实现的预测分布的点估计和不确定性估计,同时对于BDNN的实时部署来说足够快。

英文摘要:

Deep neural networks (DNNs) are known for their high prediction performance, especially in perceptual tasks such as object recognition or autonomous driving. Still, DNNs are prone to yield unreliable predictions when encountering completely new situations without indicating their uncertainty. Bayesian variants of DNNs (BDNNs), such as MC dropout BDNNs, do provide uncertainty measures. However, BDNNs are slow during test time because they rely on a sampling approach. Here we present a single shot MC dropout approximation that preserves the advantages of BDNNs without being slower than a DNN. Our approach is to analytically approximate for each layer in a fully connected network the expected value and the variance of the MC dropout signal. We evaluate our approach on different benchmark datasets and a simulated toy example. We demonstrate that our single shot MC dropout approximation resembles the point estimate and the uncertainty estimate of the predictive distribution that is achieved with an MC approach, while being fast enough for real-time deployments of BDNNs.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑