林荣荣

所属研究所、院系: 
数据科学研究所
职称: 
副研究员
E-mail: 
linrr8@mail.sysu.edu.cn
办公地点: 
超算中心大楼524室
教师简介: 

Rongrong Lin (born in 1990) is currently a Research Associate Professor of School of Data and Computer Science, Sun Yat-sen University, Guangzhou. He received Ph.D. in computational mathematics from Sun Yat-sen University, Guangzhou in June 2017. From October 2015 to October 2016, he was a Research Assistant  with the Department of Mathematical and Statistical Sciences, University of Alberta, Edmonton, Alberta, Canada.

研究领域: 

Kernel Methods in Machine Learning and Time-Frequency Analysis

代表性论著: 

1. R. Lin, H. Zhang* and J. Zhang, On Reproducing Kernel Banach Spaces: Generic Definition and Unified Framework of Constructions, 25 pages.

2. R. Lin*, An Optimal Convergence Rate for Gaussian Regularized Shannon Sampling Series, arXiv: 1711.04909, 14 pages.

3. R. Lin and H. Zhang*, Convergence Analysis of the Gaussian Regularized Shannon Sampling Series, Numerical Functional Analysis and Optimization, 38 (2017), no. 2,224–247.

4. F. Huang, R. Lin* and Y.Yang, The Circular Bedrosian Identity and Multidimensional Periodic Analytic Signals, Complex Variables and Elliptic Equations, 62(2017), no. 2, 199–213.

5. R. Lin and H. Zhang*, Existence of the Bedrosian Identity for Fourier Multiplier Operators, Forum Mathematicum, 28 (2016), no. 4, 749–759. 

6. W. Hu, R. Lin* and H. Zhang, The Circular Bedrosian Identity for Translation-Invariant Operators: Existence and Characterization, Mathematical Methods in the Applied Sciences, 38 (2015), no.18, 5264–5270.

学术会议

1. Invited talk, 25 mins. The 2nd International Conference on Kernel-based Approximation Methods in Data Analysis. Guangzhou. May 25-27, 2018.  (Topic: Reproducing Kernel Banach Spaces)

2. Invited talk, 40 mins. TSMIF Sanya workshop: From Approximation Theory to Real World Applications. TSMIF, Sanya. December 11-15, 2017. (Topic: Shannon's Sampling Theorem)

教授课程

2017.09-2018.01 Wavelet Analysis (小波分析). Outline of this course is as below.

Lecture 1: Introduction to DCT- and DWT-based JPEG

Lecture 2: Regularity of a Function and Decay of its Fourier Coefficients

Lecture 3:  DFT, FFT and DCT

Lecture 4: Fourier transform on L1(R) and L2(R)

Lecture 5: Approximation Identity and Shannon's Sampling Theorem

Lecture 6: Celebrated Results in Fourier Analysis

Lecture 7: Wavelet Analysis: Haar Wavelet

Lecture 8: General Multiresolution Analysis and the Mallat Algorithm

Lecture 9: Filter Banks (Symmetry, Vanishing Moments, Sum Rules and Linear-phase Moments)

Lecture 10: Daubechies' Orthogonal Wavelets

Lecture 11: Biorthogonal Wavelets and Discrete Wavelet Transform

Image Denoising: Matlab Code (Comparison between DCT and DWT) and Lena

 

2018.03--2018.06  An Introduction to Deep Learning (深度学习)

Lecture 1: Deep Neural Networks (DNN) & Python Code 

Lecture 2: Machine Learning Methods (k-NN,Bayes,CART,AdaBoost,SVM) & Tensorflow Code  

Project 1: Gender Identification by Voice (GitHub & Kaggle)

Lecture 3: Convolutional Neural Networks (e.g., AlexNet, GoogLeNet, VGGNet, ResNet)& Code

Assignment 1: Training DNN, SVM, CNN on the Dataset MNIST

Assignment 2: The usage of scikit-learn

Assignment 3:  MNIST (4 Conv+2 FC, 99.24% Accurary, 8 mins) & Keras Code

Lecture 4: Recurrent Neural Networks (RNN) and LSTM (Speaker: 李悦)

Lecture 5: Applications of RNNs (Precipitation Nowcasting, NLP, ASR) & Code &Word2Vector

Lecture 6: Restricted Boltzmann Machines (RBM)

Lecture 7: Generative Adversarial Networks (GAN)

Lecture 8:  Sparse Coding and Auto-Encoder (Speaker: 袁淦钊) 

Lecture 9: Object Detection (R-CNN, YOLO)

Project 2: Implementation of YOLO v3 (Code)  (Speaker: 江佳宇)

Project 3: Automatic Reading of Water Level (Image)& Keras Code&Test Image

Lecture 10: Kernel Methods for Deep Learning (RBF Network, DKL, VC Dimension)

Lecture 11: Proximal Algorithm and Majorization Minimization Algorithms (MM)

Lecture 12: Kernel Methods and Deep Networks

Talk For Undergraduates: Deep Learning and Its Application to Image Classfication

Talk For Foshan Meteorological Bureau: Deep Learning and Its Application to Meteorology