課程大綱

課程資訊

課程名稱	資料科學計算 Computation in Data Science
開課學期	109-1
授課對象	理學院應用數學科學研究所
授課教師	顏佐榕
課號	MATH5080
課程識別碼	221 U8270
班次
學分	3.0
全/半年	半年
必/選修	選修
上課時間	星期三2,3,4(9:10~12:10)
上課地點	天數305
備註	初選不開放。與潘建興、謝叔蓉合授限電資學院學生(含輔系、雙修生) 或限本系所學生(含輔系、雙修生) 總人數上限：40人
Ceiba 課程網頁	http://ceiba.ntu.edu.tw/1091MATH5080_
課程簡介影片
核心能力關聯	本課程尚未建立核心能力關連
課程大綱
為確保您我的權利,請尊重智慧財產權及不得非法影印
課程概述	本課程教授與統計以及機械學習有關的計算方法，主題包括 stochastic optimization， matrix decomposition，methods for high dimensional regression and classification，matrix differentiation，convex optimization，以及和深度學習有關的演算法。
課程目標	學生在課堂上會學到現代統計以及機械學習中常見的計算方法，例如genetic algorithms，simulated annealing， particle swarm optimization， ant colony optimization，fuzzy system optimization，Dimension reduction，Modeling with variable selection (lasso，elastic net，ridge, etc)，Prediction via ML (support vector machines，k-nearest neighbor, etc)，alternating direction methods of multipliers (ADMM)，proximal algorithms，stochastic gradient algorithms和 backpropagation等等。
課程要求	linear algebra; multivariate calculus.
預期每週課後學習時數
Office Hours	另約時間
指定閱讀	請見上課投影片。
參考書目	請見上課投影片。
評量方式 (僅供參考)

課程進度

週次	日期	單元主題
第1週	9/16	Lecturer: Frederick Kin Hing Phoa (潘建興) Lecture 1: 01-1 Introduction to Optimization 01-2 Metaheuristic Optimization: An Introduction 01-3 Traditional Methods: Random Sampling, Hill Climbing, Random Walk
第2週	9/23	Lecturer: Frederick Kin Hing Phoa (潘建興) Lecture 2: 02-1 Simulated Annealing: Metropolis Algorithm, Temperature Schedule 02-2 Tabu Search: Introduction, Iterated Local Search 02-3 Comparison between Optimization Algorithms
第3週	9/30	Lecturer: Frederick Kin Hing Phoa (潘建興) Lecture 3: 03-1 Genetic Algorithm I: Fundamental Idea and Hypothesis 03-2 Genetic Algorithm II: Selection, Crossover, Mutation 03-3 Genetic Algorithm III: Schema Theorem
第4週	10/07	Lecturer: Frederick Kin Hing Phoa (潘建興) Lecture 4: 04-1 Particle Swarm Optimization I: Basic Phenomena 04-2 Particle Swarm Optimization II: Algorithm 04-3 Particle Swarm Optimization III: Parallel Computing
第5週	10/14	Lecturer: Frederick Kin Hing Phoa (潘建興) Lecture 5: 05-1 Ant Colony Optimization I: Basic Phenomena 05-2 Ant Colony Optimization II: Algorithm 05-3 Memetic Algorithm: Basic Introduction
第6週	10/21	Lecturer: Frederick Kin Hing Phoa (潘建興) Lecture 6: 06-1 Advanced Topic I: Difficulties in Optimization 06-2 Advanced Topic II: Evolution Strategies 06-3 Final Project: A Discussion
第7週	10/28	Lecturer: Shwu-Rong Grace Shieh (謝叔蓉) Lecture 1: Introduction of Dimension Reduction Lecture 2: Principal component analysis Lecture 3: Applications to biomedical data
第8週	11/04	Lecturer: Shwu-Rong Grace Shieh (謝叔蓉) Theme: Modeling high-dimensional data Lecture 1: Introduction of Data for your projects Lecture 2: Modeling with variable selection (lasso, ridge, etc) Lecture 3: Applications to drug response prediction
第9週	11/11	Midterm week (No lecture)
第10週	11/18	Lecturer: Shwu-Rong Grace Shieh (謝叔蓉) Theme: Prediction via machine learning Lecture 1: Introduction Lecture 2: k-Nearest neighbor Lecture 3: Applications to drug response data
第11週	11/25	Lecturer: Shwu-Rong Grace Shieh (謝叔蓉) Theme: Prediction via machine learning Lecture 1: Introduction Lecture 2: Method: Support Vector Machines Lecture 3: Applications to precision medicine
第12週	12/02	Lecturer: Shwu-Rong Grace Shieh (謝叔蓉) Lecture 1: Students' presentation (5 min. to present your initial ideas for the project) Lecture 2: Students' presentation Lecture 3: Students' presentation
第13週	12/09	Lecturer: Tso-Jung Yen (顏佐榕) Matrix computation
第14週	12/16	Lecturer: Tso-Jung Yen (顏佐榕) Convex optimization I
第15週	12/23	Lecturer: Tso-Jung Yen (顏佐榕) Convex optimization II
第16週	12/30	Lecturer: Tso-Jung Yen (顏佐榕) Convex optimization III
第17週	1/06	Lecturer: Tso-Jung Yen (顏佐榕) Algorithms in deep learning