Just For My Daily Diary
Just For My Daily Diary
Introduction 简介 One of the most basic questions we might ask of a model is: What features have the biggest impact on predictions? 我们可能会问模型的一个最基本问题是：哪些特征对预测的影响最大？ This concept is called feature importance. 这个概念称为特征重要性。 There are multiple ways to measure feature importance. Some approaches answer subtly different versions of the question above. Other approaches have documented shortcomings. 有多种方法来衡量特征重要性。一些方法回答了上述问题的微妙不同版本。其他方法有记录的缺点。 […]
What Types of Insights Are Possible 可能获得哪些类型的洞察 Many people say machine learning models are "black boxes", in the sense that they can make good predictions but you can’t understand the logic behind those predictions. This statement is true in the sense that most data scientists don’t know how to extract insights from models yet. 许多人说机器学习模型是“黑匣子”，即它们可以做出正确的预测，但你无法理解这些预测背后的逻辑。这种说法是正确的，因为大多数数据科学家还不知道如何从模型中提取洞察。 […]
This notebook is an exercise in the AI Ethics course. You can reference the tutorial at this link. In the tutorial, you learned about different ways of measuring fairness of a machine learning model. In this exercise, you’ll train a few models to approve (or deny) credit card applications and analyze fairness. Don’t worry if […]
This notebook is an exercise in the Computer Vision course. You can reference the tutorial at this link. Introduction 简介 In these exercises, you’ll explore what effect various random transformations have on an image, consider what kind of augmentation might be appropriate on a given dataset, and then use data augmentation with the Car or […]
Introduction 简介 Now that you’ve learned the fundamentals of convolutional classifiers, you’re ready to move on to more advanced topics. 现在您已经了解了卷积分类器的基础知识，可以继续学习更高级的主题了。 In this lesson, you’ll learn a trick that can give a boost to your image classifiers: it’s called data augmentation. 在本课中，您将学习一种可以增强图像分类器的技巧：它被称为数据增强。 The Usefulness of Fake Data 虚假数据的用处 The best way to improve the performance […]
This notebook is an exercise in the Computer Vision course. You can reference the tutorial at this link. Introduction 介绍 In these exercises, you’ll build a custom convnet with performance competitive to the VGG16 model from Lesson 1. 在这些练习中，您将构建一个自定义卷积网络，其性能可与第 1 课中的 VGG16 模型相媲美。 Get started by running the code cell below. 通过运行下面的代码单元开始。 # Setup feedback […]
Introduction 介绍 Now that you’ve seen the layers a convnet uses to extract features, it’s time to put them together and build a network of your own! 现在您已经了解了卷积网络用于提取特征的各层，是时候将它们组合在一起并构建自己的网络了！ Simple to Refined 从简单到精细 In the last three lessons, we saw how convolutional networks perform feature extraction through three operations: filter, detect, and condense. A single round […]
This notebook is an exercise in the Computer Vision course. You can reference the tutorial at this link. Introduction 简介 In these exercises, you’ll explore the operations a couple of popular convnet architectures use for feature extraction, learn about how convnets can capture large-scale visual features through stacking layers, and finally see how convolution can […]
import numpy as np from itertools import product from skimage import draw, transform def circle(size, val=None, r_shrink=0): circle = np.zeros([size[0]+1, size[1]+1]) rr, cc = draw.circle_perimeter( size[0]//2, size[1]//2, radius=size[0]//2 – r_shrink, shape=[size[0]+1, size[1]+1], ) if val is None: circle[rr, cc] = np.random.uniform(size=circle.shape)[rr, cc] else: circle[rr, cc] = val circle = transform.resize(circle, size, order=0) return circle def […]
This notebook is an exercise in the Computer Vision course. You can reference the tutorial at this link. Introduction 介绍 In these exercises, you’ll conclude the feature extraction begun in Exercise 2, explore how invariance is created by maximum pooling, and then look at a different kind of pooling: average pooling. 在这些练习中，您将完成练习 2 中开始的特征提取，探索最大池化如何产生不变性，然后研究另一种池化：平均池化。 Run […]