There is more AI news here than can keep you connected as much as possible. However, you can stay tolerantly up-to-date on the most interesting improvements with this column, which collects AI and machine learning advances around the world and explains why they can be important to technology, startups or civilization. To begin with the shortsighted note: Researchers say the ways in which machine learning applied in fine arts are always interesting – although not always practical. A team from the University of Washington wanted to see what played on the piano and whether any computer view system could learn from the player’s overhead view.
Eli Shlizerman, Kun Su and Xiulong Liu trained the system, watching videos of playing the piano and first pulling out simple sequences like the piano-roll of the key press. It then adds expressions to the length and strength of the presses and finally polishes it to give input to a MIDI synthesizer for the output. The results are somewhat loose but certainly recognizable. Shlizerman says, “Creating music that looks like it could be played in a music performance was previously considered impossible,” “Shlizerman said. It needs to be ‘imagined’, it needs a system that is appropriate and imaginative, it was a surprise to achieve the music that we thought was so good.
Another interesting field in the field of arts and correspondence is that it is a very interesting handle are so complex and sealed. Their method was to x-ray the letters and establish a new, improved algorithm to understand the image of the result. “The algorithm does an impressive job of separating the layers of paper, despite the final thin and small gaps between them, which are sometimes less than the resolution of the scan,” said Eric Damien of MIT. “We weren’t sure it would be possible.”
The work can applied to many types of documents that are difficult to uncover with simple X-ray techniques. Classifying it as “machine learning” is a bit of an extension, but it was very interesting not to include it. Read the full paper on nature communication. You have reached a charge point for your electric car and it has gone out of service. You can even leave a bad review online. In fact, thousands of reviews exist and form a potentially highly effective map for municipalities seeking to expand the electric vehicle infrastructure. Omar Asensio of Georgia Tech trained a natural language processing model in such reviews and soon became an expert at parsing them in the thousands and perfecting insights into where outages general, comparative costs and other factors.
