機器學習之訊號處理應用(英文授課)
Machine Learning for Signal Processing
| 節 | 週一 |
|---|---|
N 12:20–13:10 | 機器學習之訊號處理應用(英文授課) ED202 3 節連堂 |
5 13:20–14:10 | |
6 14:20–15:10 |
* 根據陽明交大上課時間表所列
In this course, we would discuss the connection between signal processing and machine learning. Specifically, we would focus on applying machine learning methods for signal processing. Note that signals can be many kinds, such as wireless signals, sensor signals, images, videos, speeches. The major tasks of signal processing are to discuss the ways for extracting useful/meaningful information from signals for practical engineering problems. There are two main aspects of processing. One is the signal representation, modeling, and characterization; the other is signal categorization, estimation, and prediction. Specifically, we would discuss the learning algorithms for these signal processing aspects upon data. Students will learn contemporary techniques for processing signals, enhancing signals, classifying signals, and learning from signals. The topics include (statistic, data-driven) representation, signal detection classification and prediction, and signal modeling. With time, we may illustrate some advanced applications related to compressed sensing and deep learning in the course. After this course, you should be able to (1) Recognize and identify the technical terms of the taught learning algorithms for signal processing. (2) Explain the fundamental concept and function of these learning algorithms. (3) Apply these learning algorithms to solve the provided exercise sets. (4) Design a system to solve/provide a practical signal-processing issue/service based on some learning algorithms.
Signals and systems, Linear algebra, Probability, Programming skills
Course Lectures with exercises. TA Office: EC118 TA: Tso-Yuan Chen 陳作源(ooabcdexx@gmail.com)、 Chih-Jen Cheng 鄭智仁(king601012003.cs08g@nctu.edu.tw) Web: http://acm.cs.nctu.edu.tw
Tentative plan (1) Homework 40% (2) Challenge Exercise 20% (3) Midterm project Proposal 15% + Final Project/Paper Format Report 25%.
| 週次 | 主題 |
|---|---|
| 第 1 週 | (1)Course Introduction (2)Linear Algebra (3)Introduction of CNN |
| 第 2 週 | (1)Probability and Estimation (2) CNN architectures |
| 第 3 週 | (1)Time Signal Presentation and signal processing (2)RNNs I |
| 第 4 週 | (1) Perception and feature for Audition and vision (2) RNNs II |
| 第 5 週 | (1)Learning Feature extraction by PCA (2) Advanced CNN |
| 第 6 週 | (1)Feature extraction by ICA and Non-negative matrix factorization (2) Network training |
| 第 7 週 | (1)Non-linear dimension reduction and representation (2) Network Compression |
| 第 8 週 | (1)Non-linear dimension reduction and representation (2) Learning-based object Detection |
| 第 9 週 | Midterm |
| 第 10 週 | (1)Detection and Matching filter (2) Image segmentation |
| 第 11 週 | (1) Supervised learning (2) Similarity Learning |
| 第 12 週 | (1)Supervised learning (2) Network Visualization |
| 第 13 週 | (1)Clustering, GMMs, EM algorithms, (2) GAN |
| 第 14 週 | (1)Compressive Sensing (2) VAE |
| 第 15 週 | (1) Time-Series, HMM, DTW (2) Reinforcement Learning |
| 第 16 週 | (1)Advanced Topics |
| 第 17 週 | (1)Final Presentation |
| 第 18 週 | (1)Final Presentation |
1. “Machine Learning: A Probabilistic Perspective”, Kevin P. Murphy, MIT Press, 2012/08/24 2. “The Elements of Statistical Learning: Data Mining, Inference, and Prediction”, Trevor Hastie, Robert Tibshirani, and Jerome Friedman, Springer, 2008 3. "Machine Learning for Signal Processing Data Science, Algorithms, and Computational Statistics", Max A. Little, Oxford University Press, 2019
- 地點
- EC708
- 時間
- Tuesday, 10:00 ~ 12:00
- 聯絡方式
- chingchun@nycu.edu.tw ; chingchun@cs.nctu.edu.tw office: 54736