• En
  • /
  • Jp
  • Let’s learn about increasingly and widely utilized AI!

    • Share
    • Twitter
    • Google+
    • はてなブックマーク


     As AI (Artificial Intelligence) has been put to practical use in recent years, it is said that we are in the third AI boom. AI is evolving much faster than our expectation and its level is above our imagination.
    What exactly AI is? Let’s understand its history, practical examples in our daily life and latest technology which triggered the current AI boom.


    1-1. Overview of AI
    1-2. History and transition of AI
    1-3. Classification of AI: “Specialized type” and “General-purpose type”
    1-4. Four levels of AI
    1-5. Future of AI
    2-1. Conclusion

    Overview of AI

    AI means software or system which enables computers to accomplish intellectual works which human beings are doing. Siri installed with iPhone and robot type Pepper sold by softbank could be given as examples. Google’s AlphaGo, which has beaten the strongest professional go player the other day, might be the hottest AI at this point. In addition, AI is rapidly implemented in many other fields such as optimization of search results on search engine and stock trading.

    • siri

    • pepper

    • AlphaGo VS Lee Sedol

    History and transition of AI

    • Here, take some time to learn about the history of AI.
      The word “AI” was firstly used in late 1950’s. People started finding solutions of puzzle games and working on researches of natural language processing. This period, between late 1950’s to 1960’s is called the first AI boom.

      The second boom came in 1980’s. The research of the “expert system” development, which systematizes human knowledge on computer to help to solve problems, made progress.

      And now, we are in the third boom. Thanks to technologies such as machine learning and deep learning, the range of application of AI has dramatically extended and its precision has improved. Thus, AI has begun to be actually used for many problem-solving in the real world and this brought the third boom.

    • It introduced the 1980’s expert system
      Baseball game “Earl Weaver Baseball” released

    Classification of AI: “Specialized type” and “General-purpose type”

    There are two types of AI, “Specialized type” and “General-purpose type”.

    The “Specialized type” is AI which is developed to solve some specific tasks or to accomplish some specific goals.
    AlphaGo and AI for the support of stock trading and self-driving system are classified as this specialized type.

    The “General-purpose type” is AI which can solve many tasks that we human are faced with. Unfortunately, this general-purpose type can be seen only in SF movies for the moment.

    • Specialized type example: Autonomous driving system image

    • General-purpose type example: R2-D2 and C-3PO
      from the movie “Star Wars”

    1-4. Four levels of AI

    AI can be divided into following four levels:

    • -Level 1: Simple control program
      At this level, AI conducts predetermined reactions or operations. Temperature control of air conditioner falls under this category.

    • Practical example: Air conditioner
      temperature control system

    • -Level 2: AI with numerous corresponding patterns
      At this level, AI conducts predetermined but numerous reactions. AI geared to games such as AlphaGo, and to robotic vacuum cleaner fall under this category.

    • Practical example: Cleaning robot

    • -Level 3: AI which learns corresponding patterns automatically
      At this level, AI learns spontaneously corresponding patterns from data source by machine learning. Feature quantity, which allows AI to recognize the patterns, must be installed by human beforehand.

    • Practical example: Search engine

    • -Level 4: AI which also finds the feature quantity to learn the corresponding patterns
      At this level, AI can find the feature quantity by itself through deep learning and so on. This is the most advanced (cleverest) type of AI in existence.

    • Practical example: Image recognition

    1-5.Future of AI

    • Then, what will AI going to be in the future?
      It is anticipated that AI will continue to expand its reach of application more and more. The list of examples would be endless, but it would be applied in more critical fields such as healthcare or money and banking given the fact that “chat bot”, AI which accomplish the task using natural language or driving assistance system by emotion-recognition software, are getting practical.

      However the future of AI is not always bright. The emotional jolt delivered by an article of the Oxford is fresh in our memory. Whether it’s true or not, it is possible that AI will bring a great change in “human’s way of working”.

    2-1. Conclusion

    AI, started from finding solutions of games, has now mastered even “the ability to learn by itself”. Although it is difficult to know the future of AI as well as man’s destiny, it’s sure that AI will enlarge human’s ability.

    For the further evolution of AI, we need cultivated technologies represented by machine learning and deep learning.
    If you follow the trend of these technologies, you may get a glimpse of AI, even if you are not directly participated in the development of AI.