arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1810.04045stat.MLcs.LG
收藏

辍学作为一种结构性收缩之前

Dropout as a Structured Shrinkage Prior

University of Cambridge(剑桥大学);Microsoft Research(微软研究院);Alan Turing Institute(艾伦·图灵研究所);University of California, Irvine(加州大学欧文分校)

Eric Nalisnick, José Miguel Hernández-Lobato, Padhraic Smyth

摘要:

深度神经网络的丢弃正则化是一种神秘而有效的防止过拟合的工具。其成功的解释范围从防止“共同适应”的权重,它是一种廉价的贝叶斯推断的形式。我们提出了一个新的框架来理解神经网络中的乘性噪声,考虑了连续分布以及伯努利噪声(即辍学)。我们发现,乘性噪声诱导结构化收缩先验网络的权重。我们通过尺度混合物的重新参数化性质得到等价性,而不需要调用任何近似。鉴于等价性,我们然后表明,辍学的蒙特卡罗训练目标接近边际MAP估计。我们利用这些见解,提出了一种新的收缩框架resnets,terning之前的“自动深度确定”,因为它是网络深度的自动相关性确定的自然模拟。最后,我们研究了两种推理策略,改进了回归基准中的MAP近似。

英文摘要:

Dropout regularization of deep neural networks has been a mysterious yet effective tool to prevent overfitting. Explanations for its success range from the prevention of "co-adapted" weights to it being a form of cheap Bayesian inference. We propose a novel framework for understanding multiplicative noise in neural networks, considering continuous distributions as well as Bernoulli noise (i.e. dropout). We show that multiplicative noise induces structured shrinkage priors on a network's weights. We derive the equivalence through reparametrization properties of scale mixtures and without invoking any approximations. Given the equivalence, we then show that dropout's Monte Carlo training objective approximates marginal MAP estimation. We leverage these insights to propose a novel shrinkage framework for resnets, terming the prior 'automatic depth determination' as it is the natural analog of automatic relevance determination for network depth. Lastly, we investigate two inference strategies that improve upon the aforementioned MAP approximation in regression benchmarks.

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。

↑