Machine Learning Note: Support Vector Machine (2)

06 March 2020

4. Kernel SVM 4.1 Kernel Trick Recall the dual form of the SVM problem:

read more

Machine Learning Note: Support Vector Machine (1)

18 November 2019

1. Intuition of SVM Assume that we have some data points of two different classes which are linear separable. In the case of two dimensional data, we can use a line $y = w^Tx + b$ to separate it. If the data are high dimensional, we can use a hyperplane to separate it.

read more

Machine Learning Note: Duality

17 November 2019

1. Lower bound property Given an optimization problem with the following standard form:

read more

Paper Summary: From Word Embeddings To Document Distances

13 November 2019

Paper Title: From Word Embeddings To Document Distances

read more

Solving a simple transportation problem using CVXOPT

12 November 2019

Problem Description Assume that we are the manager of a support chain. Our company has two factory A and B, and each of them has 300 and 250 products, respectively. Now we would like to deliver them two three retail stores in different cities 1, 2, 3. There is a cost for each delivery as listed below.

read more

Machine Learning Note - Convex Optimization

11 November 2019

Introduction I’ve been taking an online Machine Learning class recently. This post is my note on convex optimization part.

read more

Training A Self-Driving Car to Learn from Human Driving Behavior

09 August 2018

I was enrolled in the Udacity Self-Driving Nanodegree starting from this summer, and this is my solution to the behavior cloning project in my Nanodegree courses. Comparing to other solution that could be found on the web, this solution is clean and simple, and the resulting model can drive the car perfectly on both tracks. The code and trained model of this project are available on [Github].

read more

Set Up A New Deep Learning Server

15 June 2018

Updates: I found that CUDA and cuDNN can be installed through Anaconda, and it has a lot of benifits (e.g install different versions of CUDA and cuDNN in different environment). So I ended up switching to this method and we don’t need to install them globally anymore.

read more