|
所在平台: Coursera |
课程主页: https://www.coursera.org/learn/advanced-ai-techniques-for-the-supply-chain
课程评论：没有评论
课程名称：供应链的高级人工智能技术 课程概述：在本课程中，我们将学习用于解决供应链问题的更高级的机器学习方法。课程开始时将概述不同的机器学习范式（回归/分类）及最新模型在这些分类中的适用性。接着，我们将深入探讨一些具体技术和应用场景，例如使用神经网络预测产品需求和利用随机森林进行产品分类。理解模型的假设和所需的预处理步骤是应用这些模型的重要部分。最后，我们将通过一个项目，结合先进技术，解决一个图像分类问题，以找出机器中产生的缺陷产品。 课程大纲： 第一部分：经典人工智能方法 描述：在本模块中，我们将涵盖与机器学习范式相关的概念。我们将学习如何选择模型，并考虑管理偏差-方差的权衡。接下来，我们将探讨机器学习模型的收敛过程，包括利用随机梯度下降法来最小化损失函数。最后，我们将讨论使用库进行超参数调整时的实际编码注意事项。 第二部分：图像与文本 描述：在本模块中，我们将超越数字数据，学习如何在图像和文本上应用机器学习。我们将首先讨论如何分析文本数据，涵盖自然语言处理的基本方法。然后，我们将学习如何通过构建卷积神经网络，分析图像，完成包含卷积和池化层的网络。 第三部分：最终项目：使用图像分类检测异常 描述：在这个最终项目中，我们将应用在最后一个模块中学到的内容，根据产品是否存在缺陷对图像进行分类。
Part: 1
Title:A Classical AI Approach
Description:In this module, we'll cover the concepts relating to the ML paradigm. We'll start by learning how to pick a model, relying on considerations such as managing the bias-variance tradeoff. Next, we'll explore how machine learning models converge, including the use of stochastic gradient descent to minimize loss functions. Finally, we'll end with some practical considerations on coding advanced AI models with libraries for hyperparamter tuning.
Part: 2
Title:Images and Text
Description:In this module, we'll expand beyond numbers and learn how to use machine learning on images and text. We'll start by talking about how to analyze text data and cover the primary methods behind natural language processing. Then, we'll learn how to analyze images by constructing convolutional neural networks complete with convolutions and pooling layers.
Part: 3
Title:Final Project: Detecting Anomalies with Image Classification
Description:In this final project, we’ll apply what we learned in the last module to classify images of products based on whether there is a defect or not.
In this course, we’ll learn about more advanced machine learning methods that are used to tackle problems in the supply chain. We’ll start with an overview of the different ML paradigms (regression/classification) and where the latest models fit into these breakdowns. Then, we’ll dive deeper into some of the specific techniques and use cases such as using neural networks to predict product demand and random forests to classify products. An important part to using these models is understanding their assumptions and required preprocessing steps. We’ll end with a project incorporating advanced techniques with an image classification problem to find faulty products coming out of a machine.
