《人工智能与医学数据计算》课程共分为十节课，首先为同学们引入人工智能与医学数据计算课程的前言知识，了解人工智能、深度学习的“前世今生”；第二节课将介绍人工智能分析目标，引入关键技术及相关概念，重点讲授人工智能两种不同技术间的区别与联系。第三、四节课介绍人工智能基本应用场景及操作环境的软硬件要求，为接下来深度学习关键技术做铺垫；第五节课将从深度学习如何架构网络来了解卷积神经网络这一关键知识点，重点了解卷积的概念；第六节课讲授深度学习与梯度下降：“一场数学求解的双向奔赴”，讲授深度学习为何使用梯度下降作为网络更新的基本方法；第七节课介绍深度学习中损失函数、激活函数、与优化器等关键知识点；在完成以上课程学习后，将进入本节课的重难点部分：深度学习反向传播算法，这部分知识点分为两节课，为同学们介绍深度学习中“反向传播”这一网络更新机制；最后，本节课以医学数据计算工程项目为实例演示如何在临床情境中通过医学数据进行深度学习运算，使同学们深入了解医学数据计算的基本流程与方法。
The course "Artificial Intelligence and Medical Data Computing" is divided into ten sessions. The first session introduces the background knowledge of artificial intelligence and medical data computing, providing an overview of the evolution of artificial intelligence and deep learning. The second session covers the objectives of artificial intelligence analysis, introducing key technologies and related concepts, with a focus on differentiating and connecting two distinct AI techniques. Sessions three and four summarize the basic application scenarios of artificial intelligence and the hardware and software requirements of operational environments, setting the stage for in-depth learning of related key technologies. Session five delves into the architecture of deep learning networks, focusing on understanding key concepts such as convolutional neural networks and the concept of convolution. Session six introduces deep learning and gradient descent, explaining why gradient descent is used as the fundamental method for network updates. Session seven introduces key concepts in deep learning such as loss functions, activation functions, and optimizers. Following these sessions, the course focuses on the challenging aspects of deep learning, particularly the backpropagation algorithm, which is split into two sessions to explain the "backward propagation" mechanism in network updates. Finally, the course concludes with a demonstration using medical data computing projects as examples to illustrate how deep learning is performed in clinical contexts, enabling students to gain a deeper understanding of the basic processes and methods of medical data computing.
