Artificial intelligence cooperating in the face of the present pandemic

This is a review of some artificial intelligence news that happened throughout the week.

One of the current problems is the detection of asymptomatic people infected with Covid19, as many of us know are people who do not have symptoms that help determine a possible case of the disease, which makes them not candidates for tests, thus avoiding detection. Researchers from MIT have discovered how to detect something that identifies them in asymptomatic people, the way they cough. For the human ear this is unnoticeable from a normal cough, but using an artificial intelligence model that analyzes recordings of forced coughing through a cell phone or laptop, something that was impossible can be achieved.

Deep learning to clean up space junk

The earth has more than 34,000 objects orbiting the earth, making them a risk mainly in space missions and artificial satellites in orbit due to the high speed with which they can be found. A deep learning algorithm is learning to measure the rotation and translation of objects in the third dimension, which is critical to capturing an orbiting object with a possibly unstable vector. The company in charge of this project is derived from EPFL together with ClearSpace. The main focus is to develop the detection of the pose of objects in 6D (3 rotations and 3 translations) in real time with the help of cameras.

Artificial intelligence solving mathematical unknowns to understand the world

Automatic learning system is helping to solve complicated differential equations, technically solves entire families of equations without the need for retraining. This helps to describe changes in bodies and liquids, movements of planets and tectonic plates, air turbulence that can affect flights and more physical phenomena of reality. These formulas are used to model space orbits, try to predict earthquakes and design airplanes. This model could help intensive computational calculations by accelerating them up to 1000 times faster, paving the way for modeling even bigger problems.

Detecting gas leaks with automatic learning

New study led by scientists at Los Alamos National Laboratory uses sensors and machine learning to locate leaks in oil and gas fields. Automating the detection of the vast natural gas network. Thanks to the fact that the system helps to detect leaks quickly, costs can be reduced since traditional detection systems require many maneuvers, regardless of how expensive and slow they are.

Artificial vision camera hampered by a referee’s bald head

Due to the current measures in the face of the present pandemic, a soccer match was held without assistants, this was broadcast in a pay-per-view event. An automated HD camera with artificial vision to track the ball mistook the ball for the bald head of a linesman, which prevented normal viewing of the game.

Thanks for reading, I hope you have informed yourself of something new. See you in the next edition.