AI looking into elusive molecules sees something scientists are missing

Involving AI in analyzing experimental data has happened before, but in this case we are talking about the MUonE project. Its creators assumed independent and accurate measurements of the muon's magnetic moment.

Read also: Physicists have wondered about this for more than 100 years. They finally predicted the melting point

As our compatriots explain in: computer scienceAI-based tools can be used to quickly reconstruct particle trajectories. If everything goes according to plan, we will see the consequences of implementing this idea in the next two or three years. It is of course about the application of technology in relation to MUonE.

During experiments performed with particle detectors, these particles pass through subsequent layers of the detector. With each such transition, they leave behind a certain amount of energy. Based on the points created in this way, scientists can reconstruct the flight path.

In the Polish scientists' concept, artificial intelligence could be used to track particles used in detectors

In theory, it seems trivial, but practice shows something completely different. Most problems occur when these points are close together, making them difficult to identify. Artificial intelligence, which is adapted to analyze many pieces of information simultaneously and is able to learn certain global patterns, could revolutionize this field.

Polish scientists are behind a so-called deep neural network in which the input layer consists of 20 neurons. The next four contain 1000 neurons, while the first contains 8 neurons. All neurons in each layer are connected to all neurons in the adjacent layer, and in total, the entire network has two million configuration parameters with values ​​determined during the learning process.

Read also: Artificial intelligence will have a future like in the famous series. Experts confirm

Then comes training time, during which the tool will learn how to reconstruct particle trajectories. It turns out that once artificial intelligence acquires the appropriate capabilities, it can perform the tasks assigned to it with high efficiency and at a rapid pace. After some time has passed, it will be time to test this concept under the difficult conditions of the MUonE experiment.

Leave a Reply

Your email address will not be published. Required fields are marked *

You May Also Like

Woodpecker. Why don’t woodpeckers get a concussion? Search

While walking in the forest, we can often hear the characteristic sounds…

Avatar recovery. This is how mankind conquered the moon

Digital photo recovery is very popular and allows us to see details…

The collision occurred at an altitude of 1,400 km above the Earth’s surface. This has become a serious problem

As Jonathan McDowell, who monitors our immediate space environment on a daily…

This is probably the strangest signal ever. And astronomers are trying to explain it.

The team that conducted the scientific investigation into this matter was led…