Tagbangers Blog


Hi, I'm Namiki.

Apparently AI is becoming increasingly popular these days.

One of the most popular ones would be IBM Watson I guess.

This might reveal my age group, but AI for me has always been Knight Rider.

・Cars would talk to their drivers.

・They can decide the destination.

・They can even fight against enemies (criminals).

・They always come when help is needed.

*This was over 20 years ago so my memories are probably not accurate.

One day I happened to wonder "What really is AI? What can they do?", so I went online. 

Then I found an interesting article ,so I'm going to share it with you.

10 common misunderstandings about AI (It's a Japanese article)

Here are some of the misunderstandings introduced in the article:

1. A highly intelligent AI already exists

I thought the machines that beat humans in chess, Shogi (Japanese chess), and Go (Chinese board game) are all AI, but it turns out they were not. 

They were just memorizing all the basic patterns and making the safe moves accordingly. So if the player (human) make too many absurd moves, the computer's processors sometimes can't catch up. However if they memorized literally all the possible patterns, we could end up in a situation where you're unable to win but still won't lose.


3. There is a, independent technology called "AI".

"Technology" seems a bit too wide for a definition; "field" might be a better term in this case.

5. "Teacher-less learning" is better than "With-teacher learning"

How about humans? If one can learn without teachers, would that mean that they are intelligent from the first place...?

8. There is an AI that can be used immediately by anybody.

I thought IBM Watson and SoftBank Pepper could be used by anybody, but I guess it's quite hard to define the term "immediate".  

10. After all, AI is useless and thus meaningless.

I'm not sure if this came up due to the limited abilities of AI we have now, or the rumors that AI would eventually replace humans, but anyways, I personally am very excited to see AI keep making progress.

Did you have any of the misunderstandings above?

Machine-Learning, Deep-Learning and AI all share the same trait; they all require effective use of Big Data.

Since these three words are so similar, I feel like their definitions and interpretations can easily get mixed up.

If that happens to you, this website (Japanese) might help.

Original article written by: Yusuke Namiki

Translated by: Yu Koga