How to apply machine learning to the universal translator?
Before we begin diving into the main topic, let’s take a step back and explain what exactly is the universal translator.
The universal translator is a software I built in python for a university class. It’s functionality it’s very simple, It takes a file with a number, the unit in which it’s expressed and the unit you want to convert it to. The program will give you back a new file with the results of the conversion. Here’s an example:
- Given file
v = value, fu = from unit and tu = to unit.
2. Result file
Now that we are on the same page, let’s go back to machine learning applications with the universal translator.
When I think of machine learning, the first thing that comes to mind it’s automation. But when I take a look at the universal translator, I don’t think there’s anything really that you can do with machine learning that would make it more useful. If we go out from it’s simple application of “translating” from one unit to another then we can start having a better conversation. Let’s look at an example to see what I mean:
Let’s imagine for a second that we have a machine that can convert some material X to another material Y (and vice versa), but to convert from X to Y you need a different amount of X material to get the amount of Y you want. Well, obviously you want all this process to be as smooth as possible so you decide to use machine learning with a variation of the universal translator. This modified version will not give you some simple conversion, It will take a blueprint of the machine you want to build, and will list you the amount of material X you will need to make a sufficient amount of material Y to make the machine. The program will know how much material you will need from past experience, and will keep “learning” as it goes from new data.
In conclusion, we can modify any simple program and make it more practical and useful if we take our imagination for a ride with the help of our friend machine learning.
