人工智慧

臺北科技大學 114-2 學期課程，課號 360004。

課程大綱

This course will focus on machine learning rather than hand-crafted knowledge engineering, due to the increased availability of data, computing resources, and new algorithms. Deep learning, probabilistic programming, and multi-agent systems receive expanded coverage, each with their own chapters. The coverage of natural language processing, robotics, and computer vision has been updated to reflect the impact of deep learning. The robotics chapter now includes robots that interact with humans and the application of reinforcement learning to robotics.

課程進度

* Introduction * Deep Learning/Machine Learning * Loss Function and Entropy * Convolutional Neural Networks (CNNs) * 1-D CNNs * Visual Geometry Group (VGG) Net * Deep Residual Learning for Image Recognition * EfficientNet * Generative Adversarial Networks (GANs) * CycleGAN * Autoencoder (AE) * Convolutional Autoencoder (CAE) * Variational Autoencoder (VAE) * Denoising Autoencoder (DAE) * U-Net for image segmentation * Vision Transformer (ViT) * Diffusion models for image generation * Diffusion models for time series prediction * Reinforcement Learning: DDPG & PPO (if time allowed)

評量方式

* Homework Assignments 80% * Participation 20% All homework assignments should be submitted individually. Students over the age of 40 may submit their assignments in pairs. 25% of Homework Assignments come from Qualcomm Edge-AI Group Term Project, including hands-on work plus the term-paper submission.

教材

【遵守智慧財產權觀念，請使用正版教科書，不得使用非法影印教科書】 使用外文原文書：是 ISBN-13: 9781292401133 Artificial Intelligence: A Modern Approach 4th Stuart Russell (Author), Peter Norvig (Author) Pearson 2022 Additional lecturing materials in PDF will be provided.