深度学习
EnEn1998
爱学习
展开
-
sigma-point阅读笔记
Sigma-point的主要内容是通过上一个sigma-point(包括状态估计和协方差)预测当前的sigma-point。sigma-point指的是状态点,测量点,噪声点,协方差点。由这些点共同预测当前时刻研究对象的状态。术语estimate:x^\hat{x}x^ 或者x_xx (下标)_{(下标)}(下标)measurement:z^\hat{z}z^ 或者y_yy (下标)_{(下标)}(下标)predicted:−减号- 减号−减号state:x^\hat{x}x^proc原创 2020-10-26 17:30:01 · 1367 阅读 · 0 评论 -
2020-09-13
这里写目录标题Units of:ϕ(x)\phi(x)ϕ(x)Blind Spots in Neural Networks二级目录三级目录Units of:ϕ(x)\phi(x)ϕ(x)x′x'x′为图像 I\IotaI为x′=argmax⏟x∈I⟨ϕ(x),ei⟩x'=\underbrace{arg max}_{x\in \Iota} \langle \phi(x),e_i \ranglex′=x∈Iargmax⟨ϕ(x),ei⟩Blind Spots in Neural Network原创 2020-09-22 09:21:57 · 173 阅读 · 0 评论 -
EXPLAINING AND HARNESSING ADVERSARIAL EXAMPLES
阅读笔记 EXPLAINING AND HARNESSING ADVERSARIAL EXAMPLESABSTRACTINTRODUCTIONRELATED WORKTHE LINEAR EXPLANATION OF ADVERSARIAL EXAMPLESLINEAR PERTURBATION OF NON-LINEAR MODELSADVERSARIAL TRAINING OF LINEAR MODELS VERSUS WEIGHT DECAY问题初次编辑于2020.9.7 by EnEnABSTR原创 2020-09-07 19:11:35 · 185 阅读 · 0 评论 -
阅读笔记 DeepFool: a simple and accurate method to fool deep neural networks
阅读笔记 DeepFool: a simple and accurate method to fool deep neural networksAbstractIntroductionDeepFool for binary classifiersDeepFool for multiclass classifiersAffine multiclass classifierGeneral classifierExtension to lpl_plp normAbstract提出Deep Fool 算法原创 2020-09-01 17:49:01 · 380 阅读 · 1 评论 -
Neural Networks and Deep Learning阅读笔记(第六章第三节)
阅读笔记Recent progress in image recognitionThe 2012 LRMD paperThe 2012 KSH paperThe 2014 ILSVRC competitionOther activityOther approaches to deep neural netsRecurrent neural networks (RNNs)Long short-term memory units (LSTMs)Deep belief nets, generative model原创 2020-08-24 21:42:29 · 163 阅读 · 0 评论 -
Neural Networks and Deep Learning阅读笔记(第六章第二节)
本文根据neural networks and deep learning一书第六章,写下总结和阅读笔记原创 2020-08-17 14:13:33 · 231 阅读 · 0 评论