jskdr

[jskdr]

I have used Tumblr for a long time ago

I am not using it now. But I will use it more. One of the advantages of this is a property that allows users to write more characters than Twitter.

[jskdr]

Speed of Octave can be enhanced by the use of C++ code with dedicated pre-complier, MKOCTFILE!

[jskdr]

Long time no come. However, I love this SNS.

[jskdr]

I like to use this since I like it. I want to discuss math here. I am not good at it. However, I wanna discuss about it together. I want to prove that math and coding are not the different but pursuade the same goal. When \(a \ne 0\), there are two solutions to \(ax^2 + bx + c = 0\) and they are \[x = {-b \pm \sqrt{b^2-4ac} \over 2a}.\]

[jskdr]

Import AI 146: Making art with CycleGANs; Google and ARM team-up on low-power ML; and deliberately designing AI for scary uses

Import AI 146: Making art with CycleGANs; Google and ARM team-up on low-power ML; and deliberately designing AI for scary uses

[jskdr]

I found that there are many artists here. I wanted be an artist when I was young. Now, I am an engineer but I desired to be a great artist like Picaso. My role models are Picaso and Einestein. I wanna be one of them when I was young. I read their stories and it touched my mind.

[jskdr]

How can I link Tumblr and Instagram.

[jskdr]

Long time no use this Tumblr. I heard that this is one of the convenient tool for SNS. This looks a blog but also works as SNS like others. I am using Facebook, Linked-in. Sometimes, I use Twitter. Now I am trying to use Instagram since young people are using it. I will try to use Tumblr more often since it is also convenit and powerful.

[jskdr]

I like it

[jskdr]

텀블러는 재밌는 SNS라 생각된다

[jskdr]

print("Hello") for i in range(10): print(i)

[jskdr]

This is a piece of computer code:

print("Hello")

I wrote code in Python. Then the output becomes:

Hello

[jskdr]

This is the perfect fomulation to represent a neural network in each layer,

Since the number of nodes in each layer can be varied, we can rewrite the formulation as:

$$x^{j+1} = f^{j} \left( \sum_{i=1}^{N^j} w_i^{j} x_i^{j} + b^j \right)$$

where \(N^j\) is the number of nodes at layer \(j\).

[jskdr]

Since the number of nodes in each layer can be varied, we can rewrite the formulation as:

$$x^{j+1} = f^{j} \left( \sum_{i=1}^{N^j} w_i^{j} x_i^{j} + b^j \right)$$

where \(N^j\) is the number of nodes at layer \(j\).

[jskdr]

Then, outputs of the last layer go to the next layer.

[jskdr]

In the neural networks, the output of a lyaer can be an input of the next layer. Hence, the output becomes repreated as follows.

$$x^{j+1} = f^{j} \left( \sum_{i=1}^{n} w_i^{j} x_i^{j} + b^j \right)$$

[jskdr]

The neural networks rely on activiation function. Then the output of the layer can be represented by the use of an activation function. Hence, the ouput is given by

$$y = f \left( \sum_{i=1}^{n} w_i x_i + b \right)$$

where \(f()\) is an activation function. The activation function can be a sigmoid, tanh, or any functions.