I'm new to programming.

I want to learn about artificial intelligence.
What language should I learn and how?

A beginner who has only done ExcelVBA.
Thank you.

  • Answer # 1

    If you want to learn the theory about artificial intelligence, you don't need a programming language, you can start with knowledge of mathematics. I don't have much interest in hard theory, but if you want to make an application like that, I think you should learn Python with abundant scientific and technological libraries. There are official tutorials on the web, and if you search on Amazon etc., there are many books that introduce implementations of artificial intelligence applications in Python (although it is mostly foreign books).

  • Answer # 2

    A person running an artificial intelligence venture.

    If you're going to learn artificial intelligence, I think it's better to work on theory and practice in parallel. Reading theory is a good way to read a book, and practicing is to create a program so that you can imagine how it works.

    Study of theory

    The most famous and readable book in recent days is
    Does artificial intelligence surpass humans?
    I think.


    First, let's study Python.
    I think Python is the most widely used for artificial intelligence programming.
    Although you study, the basic concepts of variables, functions, iteration, conditional branching, etc. are not much different from Excel VBA. The only difference is the grammar, that is, how to write.

    Because you've been studying Python too much, you can't help but understand the basic grammar. Next, you should actually write an artificial intelligence program. Python has a library called Scikit-learn. With Scikit-learn, you can actually implement artificial intelligence (machine learning) in just a few lines.

    At first, I think that it would be better to implement a linear regression (the simplest machine learning method) using Scikit-learn. It is easier to understand if you start from imitating with reference to Qiita's article (regression analysis with Scikit-learn).

  • Answer # 3

    C ++ is a lot for robot control. Also, in terms of machine learning, researchers are using MATLAB. However, C ++ takes time to learn programming, and MATLAB is a fee, so the first start is a high hurdle.

    If you want to get an overview of artificial intelligence, I recommend that you first touch the various machine learning libraries in Python to get a feel for what is possible and what is difficult. Is the learning method. Python has a machine learning library called scikit-learn. If you want to buy a book, you should buy a book with a sample program of scikit-learn (for example, "Python machine learning programming"). A machine learning library is also included in the image processing library called OpenCV. The recommended Python book here is "Practical OpenCV for Python".

    I wrote that C ++ has a high hurdle, but there are many free libraries here, there is also a C ++ library of OpenCV, and "mlpack" is also famous.

    Machine learning books have lots of mathematical formulas, but if you don't aim to develop new algorithms, you'll use a library and don't need to fully understand the formulas at first. I think it would be nice to skip the mathematical formulas for the time being and move the sample program in various ways to understand the outline.