Deep Learning starting from zero

perceptron: receive input(s) and return a single output

e.g.) y = x1*w1 + x2*w2 + b

create AND, NAND, OR, XOR with Python

Activation functions

Neuron == Node in the book

Sigmoid function

Relu (Rectified Linear Unit) function

  • Machine Learning
    • Classification 分類問題: Softmax function
      • e.g.) Picture of people -> Man or Woman
    • Regression 回帰問題: Identify function
      • e.g.) Picture of people -> The person’s weight

SQL Anti-Pattern

13. Fear of the Unknown

NULL is not falsy value

NULL is not equal to

  • false
  • 0
  • ””

Unlike usual programming language’s 2 values logic, SQL is 3 values logic. (TRUE, FALSE, and NULL)

SELECT * FROM Bugs WHERE assigned_to <> 123;
-- Not includes Bugs.assigned_to IS NULL rows

-- You have to
SELECT * FROM Bugs WHERE assigned_to <> 123 OR assigned_to IS NULL;
-- or
SELECT * FROM Bugs WHERE assigned_to IS DISTINCT FROM 123;

Go

A Tour of Go / Youtube

Interface

  • Writer / Reader
  • Stringer
  • etc.

Reflect

型情報

Concurrency

  • Concurrency … deal with multiple things simultaneously
  • Parallelism … run multiple things simultaneously

fmt

show struct’s value in different way

type Lang struct {
    Name string
    Year int
}

lang := Lang{"Golang", "2004"}

fmt.Printf("%v\n", lang)
fmt.Printf("%+v\n", lang)
fmt.Printf("%#v\n", lang) // more precise

slice

// 2nd arg means length of the slice
// 3rd arg means capacity of the slice
s := make([]int, 10, 15)


// slice which is just declared equals to nil
var s []int

if s == nil {
    fmt.Println("Wow!")
}