本課程旨在給學生對於人工神經網路的理論基礎和應用已經在軟硬體上的實現有概略的介紹，可以視為machine 
<br>learning 的進階課程，它將討論典型的主題，包括從統計學習理論(statistical learning theory)的觀點，嚴謹的討論學
<br>習的統計基礎，例如神經網絡的訓練(training)和測試(testing)、VC 維度(VC dimension)和模型複雜性(model 
<br>complexity)、正則化(regularization)作為過度擬合(overfitting)的補救措施。此外，我們還想介紹許多特定的神經網
<br>絡模型，其中一些植根於統計物理學，一些植根於生物學。例如 Hopfield 模型、玻爾茲曼機(Boltzmann machine)、
<br>脈衝神經網絡(spiking neural network)和自組織(self-organization mapping); 會以Aurelien Geron 的著作， 
<br>Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow, 作為給學生實際寫code 的實例。以及讓
<br>學生了解如何用keras等軟體實行deep neural network ，也會有大量deep learning 的實例。
<br>
<br>
<br>課程內容：
<br>1.Introduction 
<br>2. Learning process
<br>3. The perceptron
<br>4. Least-mean square algorithm
<br>5. Multil<x>ayer perceptron
<br>6. Deep neural network and its software implementation
<br>7. Convolutional neural network;
<br>8.Recurrent networks 
<br>9. Neurodynamics and Spiking neural network
<br>10. Neural network rooted in statistical physics
<br>11. Implementation in circuits and hardware
<br>
<br>
<br>擬用教科書或參考書：
<br>
<br>教科書： Aurelien Geron, Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow, 2nd Edition, 
<br>OReilly, 2019.
<br>
<br>參考書：
<br>1. Ian&#160;Goodfellow&#160;and Yoshua Bengio and Aaron Courville, Deep learning, MIT Press.
<br>2. John A. Herz, Introduction to the theory of neural computation (Santa Fe Institute Series)
<br>3. Yaser Abu-Monstafa Learning from data, AML book.
<br>4. Simon Haykin, Neural network and learning machines 3rd edition
<br>