Google aims to apply its machine learning engine ‘TensorFlow’ to determine whether or not artificial Intelligence systems can be programmed and trained to create music and videos on their own.
During the Moogfest musical and technology festival held in North Carolina, Alphabet Inc.’s (NADAQ:GOOGL) artificial intelligence researcher, Douglas Eck, stated that the company would collaborate with a new group to figure out if artificial intelligence can truly be creative and produce on its own. The audience at the Moogfest was given a brief about what can be expected from the highly promising collaboration.
The collaboration involves Magenta utilizing Google’s machine learning engine ‘TensorFlow’, enabling artificial intelligence systems to cope with the task of independently creating on their own. The main aim of the collaboration is to design working software that can train artificial intelligence systems to create musical compositions and artistic videos. This will be a major coup if the collaboration successfully accomplishes this task due to the technicalities involved.
Google is well aware that even the most advanced artificial intelligence systems cannot completely imitate the styles and techniques of existing artists and musicians. So, it’s a huge task to work on system that enables artificial intelligence software to produce entirely original ideas on its own.
The goal of the collaboration is to develop an artificial intelligence system that can independently create “musical chills” by implementing new musical pieces on a regular basis while the users sit back and relax in the comfort of their homes. The software is still in its early development stages, so might require some sort of human involvement to function. The task of programming a machine to be completely independent is extremely difficult, so the program will need some time to achieve the collaboration’s goal and evolve accordingly.
Google will start developing software tools to help their own team and other researchers to determine the creative potential of artificial intelligence systems. The first step for Google is to develop and launch a new program that assists researchers to transfer data from music files into TensorFlow, allowing the system to learn more about the musical styles and harness knowledge.
The inspiration for the new project comes from Google’s brain project ‘Google DeepDream’, where artificial intelligence systems were trained to fill the missing gaps in pictures and complete image databases. From there, Google wanted to know if it could stretch its capabilities and collaborate to create an artificial intelligence system that can make exciting music. While the idea of machines replacing chart topping musical artists is exciting and stretches the imagination of a developer, this is highly unlikely to be achieved anytime soon. Google believes in this project and aims to break the boundaries set for artificial intelligence.