News of the week selected by Impactscool – September 7th
7 September 2020 | Written by 1MP4CTSC00L
During the presentation on the night of August 28, Elon Musk unveiled the new goals achieved by Neuralink, the startup that aims to create a system capable of connecting the human brain with a machine. The progress made since the last presentation is concrete but still very far from the startup’s objectives. The most advanced form of the device was presented, first only on paper, now not only in flesh and blood, or rather, plastic and circuits, but also implanted in some pigs and capable of picking up their brain impulses. The device is called “Link” and is a kind of “fitbit in the skull”. It will be installed using a special surgical machine that in less than an hour under local anesthesia would be able to puncture a sention of the skull the size of a coin and replace it with the “Link”. The device, which then disappears under the scalp, is able to transmit the collected brain impulses through 1024 very thin cables that are implanted directly into the cerebral cortex.
Discovering new exoplanets, that is, planets that do not orbit our star, is no small task. This is a massive effort of analyzing data of various kinds, from variations in brightness to micro-shifts of stars, which astronomers are involved in full time, sometimes even asking for help from non-researchers. But this type of activity is the daily bread for artificial intelligence systems that exploit machine learning to analyze huge amounts of data in search of patterns and patterns, such as, for example, the cyclical darkening of a star by a planet that orbit in front of it. A research team from Warwick University fed its algorithm with the huge data archive collected by the Kepler and TESS missions and the result was 50 new planets, previously unnoticed under the nose of researchers. AI needs to be perfected, in a process of constant improvement, but there is no doubt that in the near future it will help astronomers to identify more and more exoplanets, speeding up and automating the process.
For years, scientists have been trying to develop a system to generate holograms and in part, in the past, they have succeeded, but always with particular optical tricks or using special three-dimensional “screens”. Today, a group of researchers from Tokyo University of Agriculture and Technology has developed a hologram that comes very close to those seen in science fiction films. To do this, the researchers used a nanomaterial that has microscopic flakes, smaller than the wavelength of the laser used to project the image of the demonstration (a rotating globe). The creation of this material is very expensive: 800 hours of preparation are required to create a 6-minute hologram. Now researchers at the Japanese university are working to perfect the system and to show color images.
Microsoft has announced the launch of a new tool called Microsoft Video Authenticator that can analyze photos and videos and provide a percentage of how much that media has been altered through deepfakes. Deepfakes are videos in which a person’s face is replaced with that of another, thus making them appear in unwanted contexts or making them do or say things that that person would not do or say. A powerful tool, especially in the hands of those who want to manipulate information, a significant risk with the arrival, in the United States, of the 2020 elections. Today a careful eye is able to find those small imperfections that reveal manipulation, but by the very nature of deepfakes, which exploit machine learning algorithms to make videos increasingly indistinguishable from unaltered videos, a tool like the one developed by Microsoft could prove very useful.