一、課程說明(Course Description)
<br>近年來因《深度學習》技術突破的緣故，
<br>類神經網路(Neural Networks, NN)這個領域重新翻紅，
<br>在語音辨識、電腦視覺等領域據說有超越傳統技術的表現。
<br>本課程將綜合數學推導、電腦作業、文獻選讀三種學習方法，
<br>期望教學相長，帶領各位同學一同進入NN 的堂奧。
<br>
<br>Prerequisites: 微積分、線性代數、機率、程式設計。
<br>
<br>相關科目或領域但非pre-requisite：神經科學、Machine Learning, Statistical mechanics.
<br>
<br>二、指定用書(Text Books)
<br>無
<br>
<br>三、參考書籍(References)
<br>1. S. Haykin, Neural Networks: A Comprehensive Foundation, 2nd Ed. Prentice Hall, 1999.
<br>註: 此書堪稱經典，已有第三版，但價格昂貴。
<br>考慮到第二版基礎的章節應仍然適用，
<br>我目前傾向沿用此書作為主要教學依據。
<br>此書網路上有 pdf 版本。
<br>
<br>2. This coursera course looks great
<br>https://www.coursera.org/course/neuralnets
<br>
<br>3. G. Hinton, S. Osindero, Y.-W. Teh (2006) "A fast learning algorithm for deep belief nets," 
<br>Neural Computation 18:1527-1554.
<br>
<br>4. Y. LeCun et al. (1989). "Backpropagation applied to handwritten zip code recognition," 
<br>Neural Computation 1:541-551.
<br>
<br>5. G. Hinton, L. deng, et al. (2012). "Deep neural networks for acoustic modeling in speech recognition," 
<br>IEEE Signal Proc. Mag. Nov. 2012, pp. 82-97
<br>
<br>
<br>四、教學方式(Teaching Method)
<br>每週兩小時lectures, 一小時實作或討論課
<br>
<br>五、教學進度(Syllabus)[暫定]
<br>Since this will be the first time the lecturer offers this course, 
<br>the syllabus is not finalized yet (as of end of Dec. 2014). 
<br>Nevertheless, lectures and homework will very likely cover these topics:
<br>
<br>1. Adaptive linear signal processing: gradient-descent methods
<br>2. Single-layer perceptron
<br>3. Multiple-layer perceptrons and the back propagation method
<br>4. Restricted Boltzmann Machines
<br>5. Dynamics of biological neural networks
<br>
<br>六、成績考核(Evaluation)
<br>100% course participation and homework
<br>
<br>
<br>七、可連結之網頁位址
<br>本課程教材、進度與內容將定期在 LMS 系統更新