[GAME REVIEW] while True: learn() - it talks all about machine learning, big data and AI.

Hey guys, this is Vivy here, thank you so much for  stopping by my channel. Today I'm gonna introduce   an educational puzzle game called while true:  learn. The name already looks very geek, through   this game, you will learn about the history of  machine learning, big data, AI, and… some very   fundamental programming knowledge and logic.

The features in the game are cats, modules,   programming, machine learning, AI..They combine  together in a very smart and interesting way.  Today, let’s see how professional  we can be after playing this game. So the game starts from, the player  wants to understand what his cat is   speaking.

Then he went on the website  to ask, and some guru told him some   information and passed on some work. And the game starts from here. To understand the cats, you will need  to be able to recognize the cat's shape,   learn their facial expressions, get more cats  involved so you will have more samples… just   like what we do nowadays to do machine learning  things.

And basically it is talking about the   fundamentals of machine learning, then  neural networks, big data and lastly AI,   it talks about the history of the whole thing  as early as 1935, when everything started,   to the AI as the current date. Yeah, this game  is serious, not joking around.

Some information   it provides are very basic but I never heard  of them before. Very interesting to know that. Each level is a job that is referred from  someone. You can tell those jobs are fake,   since the names, logic are made up, and there  are some puns, some silly stories, mocking like   rescue from the mars, helping the FBI, delivery  company, tarari.

Different things… but in some   way, they also look realistic, you can see  some ideas through them, and you can tell it   could totally happen in the real world. Except  for the jobs, you can use your knowledge to   invest in a startup, build the algorithm and  let it run automatically, then just get money   day by day..

Well there is a chance that you  will get bankrupt if you don’t manage it well. So here is how to play. In each level,  you will need to assemble the modules,   to sort and filter the data on the left, which has  three colors and three shapes, according to the   output requirements on the right side.

The modules  are different algorithms, the game will introduce   the modules as the story moves on. It coversthe  very early and basic one, the decision tree,   then the regression tree, then to MARK II, then  to neutral networks, deep learning, then to the   AI.

Basically I guess every essential one for  machine learning. And they offer related articles   and videos for you to explore on your own. When I was playing, I partly feel it’s   very gamey, like those puzzle, logic  games. But when I got stuck sometimes,   and I went to search for some related knowledge  about the algorithms, and tried them in the game,   it’s a little difficult to relate the game  with real algorithms, but somehow it worked.

Here is an example using some “geek” language.  What I learn from searching is that the algorithm   module decision tree, which is the most basic  one we use most in the game, in the real world it   will do a yes or no decision. But in the game it  doesn’t work like that.

The color decision tree,   if you choose to separate red and blue, the  red output will get red and green, the blue   output will get blue and green, it’s not a yes  or no answer. So it’s a little tricky to relate   the real with the game. And the final output  always requires different colors and shapes,   it’s not a simple yes or no situation.

But  in someway, it made me feel very realistic,   because in the real world, we don’t always  need a yes or no, either this or that result,   the programmers just need to find a way to  apply the simple yes or no decision tree   to achieve the needs, to get the  more complicated various results.

Another realistic thing in the game is, you need  to beat the time and use modules within a limit   to get the gold medal. I feel that I did need to  observe the original data set, and calculate the   processing time, and after some searchings I will  consider the GINI to calculate the efficiency,   and try to optimize the process.

I was arranging  the modules, and did testing all the time, just   like what a programmer will do in the real life.  It’s tricky to apply the knowledge from the real   world in the game, since it looks so different,  but surprisingly I got it worked.

I hope it’s not   just because I’m lucky. I hope it’s because I  understood how the developers set the mechanism. And…. another another realistic thing is, the  money in the game, you can use it to upgrade   hardwares, to improve the data processing speed,  increase the capacity, and so.

Which is just like   what we do in the real world. They have puns  everywhere. For some early levels, sometimes   you just can’t beat the time, after upgrading  the equipment, you just run the program again,   and you could get the gold medal easily.

And the  balancer in the game could separate the data,   which is like the server in real world,  that you need to rent. Ennn, you could   also buy some decos that do nothing but  entertainment and increase your happiness. Another thing that impressed me a lot in the game  is, it made me think about reality.

Sometimes you   don’t need to get a high accuracy, or you need  to get everything correct. I will consider the   samples needed, and processing time, and server  rental fee, that eventually i might decide to   abandon some data. Such as the candy factory,  if it’s in the real world, abandon the data,   means throw away the candys, that means waste.

But  by abandoning some candies it will increase the   speed and get a higher accuracy. So, I’m thinking,  maybe this is where the waste and processing fee   come from. A better algorithm or equipment might  be able to help with it. Technology changes life.

If you only play the game, it is a short  game. It took me about a week or so   to finish, since I played and learned  from other sources at the same time.   I felt I’m part of this field after playing this  game. For sure I can’t become a machine learning   guru after playing the game, but this high tech  thing doesn’t look that far away to approach.   It’s a very nice tool to help you step into this  field, and it not only just bring you there,   it also leads you to some deeper learnings,  by offering many resources.

And this is super   helpful and thoughtful. So if you would like,  you could directly step forward after this game. this game is on most main platforms, I bought  the iPad version, which i don’t recommend.   I have some troubles dragging and  selecting.

I assume the switch version   might have similar issues. So get a steam  version to play on PC on Mac if you could.   The prices are different,  ipad version is 5 dollars,   and switch and steam version is 13 dollars. Maybe  there will be some more functions and content? I really love the concept of the game, it’s  for adults, but still using a format as a game,   to turn the boring learning to some interesting  knowledge.

The mechanism is rather addictive   and the dialogs and puns always made me  laugh. It is not as hard as a hardcore game.   it’s helpful for those who want to start  learning about machine learning, since it   has so many related links. to those who just  want to generally know about machine learning,   the game does contain enough information.

I hope there will be more games like these,   that will make you learn while playing, that  will make people think. I will keep trying to   find these kinds of games and introduce them  to you. If you have some recommendations,   please comment below, share  with me and other audiences.