Output “art”, looks beautiful

  In the past few years, with the rapid development of artificial intelligence, many technology companies, research institutions and individuals are exploring the use of AI (artificial intelligence) for artistic creation. Inputting data and outputting “art” seems to have become a special product of the development of the times.
  In the process of exploring AI, various technology giants have played an important role. As early as 2015, Google launched the artificial intelligence system Deep Dream, recognizing images and repainting. In 2016, IBM’s artificial intelligence Watson completed the editing of the movie “Morgan” trailer. In the same year, Sony’s artificial intelligence Flow Machines created a pop song with the Beatles style. Also in 2016, the science fiction novel “The Day the Computer Writes Novels” created by artificial intelligence developed in Japan deceived all human judges, and was successfully shortlisted for the Japan Miniature Literature Award. In 2017, MIT created Shelley, an AI system capable of creating horror novels.
  From participating in production to independent creation, artificial intelligence seems to be going further and further on the road of art.
  But think about it, isn’t creativity unique to humans? Could art be the next field of artificial intelligence “take over”?
  As a type of artificial intelligence, neural networks known as “simulating the human brain” have some creative advantages. They are good at training large data sets, identifying patterns in these data sets, and producing outputs that follow the same rules. Neural networks can “listen” to Bach and Mozart’s music in a few hours, and they can also train and learn from Shakespeare’s works to create works that are “cookable”. Artificially created ideas have become so common that some people have forcibly trained neural network “robots” on human text samples.
  The novelist Ross Goodwin made such a bold attempt. He took a road trip across the United States-looking for something important on the journey and writing down his experience. However, the “it” that wrote the novel is just a microphone, a GPS, a camera and a laptop.
  Once the neural network is trained, “it” can produce any length of text the author wants (whether it is random or generated from a specific seed word or phrase). Ross Goodwin used the sights and sounds of road trips to provide these seeds: the novel is based on images, places, conversations with microphones, and even the clock inside the computer, writing only one sentence at a time.
  The result was mixed.
  The novel uses a proper method to complete the beginning of the content, citing the concept of time: “It was nine seventeen in the morning, and the house was heavy (9:17 in the morning, the house is very heavy).” According to the input to the algorithm In the Foursquare data set, AI began to describe the location, but soon deviated from the theme and became surreal. Although the literary experiment is a wonderful thing, repeatedly quoting the latitude and longitude coordinates will obviously not bring the reader a better reading experience.
  When talking about his novels created with machines, Ross Goodwin said: “This is a very imperfect document, just a rapid prototyping project, and the output is not perfect. I don’t think this is a human novel, or any Things related to novels. ”
  Art is a unique way of expressing human thoughts: romance or heroism, sorrow or joy, pain or joy, anger or caring … Each piece of art represents the author’s unique experience and mood. Through these works, we are like talking to the author. However, when AI strikes, we will suddenly face these questions: Do we need to redefine “art”? Do AI’s works bring us the same resonance as artists?
   But one thing is certain, even when works created by AI are popular, we don’t need to worry about the way out for human artists. At that time, real artists will only appear more precious. Because their experiences, personalities, thoughts, and emotions can never be replaced by AI.