[PYTHON] Impressions of receiving a long-term AI seminar (for those who are uncertain about attending seminars, etc.)

As the title says, I have received a long-term seminar at the URL below. As a result, it was a very good content to take, so I would like to convey it as information to people who are at a loss.

I received "a 6-month long-term course to become a self-propelled AI human resource" https://seminar.kikagaku.co.jp/

First the result. What happened in the end?

Finally, you can do the following.

    1. Machine Learning (including Deep learning) skill improvement With kaggle's Titanic exercises, I can now get a Score of about 0.80. I also participate in competitions little by little.   https://www.kaggle.com/c/titanic
  1. Acquiring comprehensive technology from scraping to web application creation Creation of loan examination application, animal classification application, etc. in the course. I think I was trained a lot in that. In addition, I actually made an app like the one at the URL below (still being improved).

○ Create a homepage for an acquaintance on Heroku (with feeder function)   https://yupi-profile.herokuapp.com/

○ Create a line art application with Google App Engine   https://aijunk.net (Stopped due to various reasons)

Background to the course

What was the state (skill) in the first place?

In my own experience, when I was a student, I was just touching C ++ in physics calculations.

I couldn't do it by myself and was frustrated Short-term courses are useless, video services do not continue, etc. I just couldn't recommend it myself. It's a matter of my personality, but how to continue for a long time Judging that you have to pay money and force it It is the background that I decided to take this course.

Why did you choose this course?

Based on the above assumptions Set as follows as a condition.

    1. It is assumed that you will attend for a long time (3 to 6 months).
  1. Being able to implement even WEB applications.
    1. Since I will be working, the class must be held on either Saturday or Sunday.

I thought it was money "Fourth Industrial Revolution Skill Acquisition Course Certification System" There was.

Fourth Industrial Revolution Skill Acquisition Course Certification System

image.png

It is a system that the government supports for courses such as AI, IoT, and advanced security networks. Click here for details. https://www.meti.go.jp/policy/economy/jinzai/reskillprograms/index.html

If you are certified as a professional practice education and training benefit by this course certification The system will support the ** up to 70% tuition fee **. Actually, it's best to ask the person in charge of the course or Hello Work. It seems that you may or may not be able to receive it depending on your employment insurance situation. Details can be found here. https://www.mhlw.go.jp/stf/seisakunitsuite/bunya/koyou_roudou/jinzaikaihatsu/kyouiku.html You can also search the list of certified courses here. https://www.kyufu.mhlw.go.jp/kensaku/SSR/SSR101Scr01S/SSR101Scr01SInit.form (Be careful because there are general and specialized.)

Finally narrow down to two.

Fortunately I met the conditions of the benefit I narrowed down to Datamix and Kikagaku.

The difference between the two companies is that I, a beginner, didn't really understand. Rather than wanting to be a data scientist or a particular profession I wanted to create an AI app **, so I chose Kikagaku-san. If you are a data scientist or specialized in analysis Datamix may be better.

For Kikagaku, it was also attractive to be able to take the E qualification. Also, I once watched Kikagaku's video on the video service. It may have been great that it was very easy to understand.

Course content

[Before attending]

It will be a form of pre-learning Udemy videos. I learned the basic knowledge of artificial intelligence and the basic syntax of python here.

[1st to 12th week]

It feels like the instructor actually copies the sutras on the screen while actually copying them. You can also refer to Kikagaku's original materials, You can also watch the video at a later date. Questions can be asked in Slack You can save weekday questions and ask questions by going early before taking the course.

It was very easy to write the math part in a notebook because it was Kikagaku style. If you want to get a feel for it, you can see it by watching this Udemy video (for a fee). https://www.udemy.com/course/kikagaku_blackbox_1/

In addition, there are the following similar materials as reference materials. The format looks like this (the actual difficulty is a little easier). https://japan-medical-ai.github.io/medical-ai-course-materials/ I think the following will be helpful for the difficulty level and how to proceed. https://tutorials.chainer.org/ja/

Both seem to be cooperation or created sites created by Mr. Kikagaku.

The main lessons learned for each period are as follows.

Weeks Contents keyword
1~3 Environment construction, regression analysis, scraping, web application numpy, pandas, jupyter, scikit-learn, Heroku, Flask, selenium, html, css, git
4~6 Deep learning, environment construction, natural language processing chainer, docker, google colaboratory, OpenCV, Pillow, MeCab
7~9 Loan screening app creation django, docker-compose
10~12 Image classification application creation All skills so far

[Weeks 13-24]

Based on your skills so far, you will be able to create the app you want to create yourself. First, consult with the instructor, tell only the image, and then make your own. If you don't understand it, it's like asking the instructor. I make it while going every week, so it was good to have an environment where I could do it even if I didn't like it.

By the way, I had a Saturday course, but it was possible to look back on Sunday. It's a small thing, but it's good that it responds quite flexibly.

Everyone was making various apps and it was encouraging.

Finally

I think it can be said for all studies ** I think that maintaining motivation is the most important and difficult **. I think the reason I've been frustrated so far is because I couldn't keep up my motivation. In this era when self-study tools such as e-learning are commonplace It may not match Making things with people who have the same purpose every week Known as one way to stay motivated I thought it was a very good experience.

For those who do not have the courage to take a step We hope you find this information useful.

Thank you for reading until the end.

Recommended Posts

Impressions of receiving a long-term AI seminar (for those who are uncertain about attending seminars, etc.)
A story about an engineer who came only on the server side created a portfolio
ABC's A problem analysis for the past 15 times to send to those who are new to Python
Impressions of receiving a long-term AI seminar (for those who are uncertain about attending seminars, etc.)
[Solved] I have a question for those who are familiar with Python mechanize.
Dart grammar for those who only know Python
A story about running Python on PHP on Heroku
For Debian users who are having trouble with a bug in the Linux kernel 5.10
Impressions of using Flask for a month
[Solved] I have a question for those who are familiar with Python mechanize.