Changing the rules of the game for future collider experiments, the Chinese Academy of Sciences and Peking University develop AI-driven injection source identification technology

Schematic diagram of a dual injection event simulated in the Circular Electron Positron Collider (CEPC). (Source: Institute of High Energy Physics, Chinese Academy of Sciences)

Editor|Ziluo

Quarks and gluons are elementary particles in the standard model of particle physics. Unlike electrons or photons, quarks and gluons cannot move freely in space and time and can only Can be bound into composite particles such as protons or neutrons.

The high-energy quarks or gluons produced in the collider experiment will be converted into a large number of final particles through complex interactions, which will be ejected at a smaller angle along the direction of motion of the original quarks or gluons. , this phenomenon is called injection.

Recently, Ruan Manqi’s team from the Institute of High Energy Physics, Chinese Academy of Sciences, Zhou Chen’s team from Peking University, and researcher Qu Huilin from CERN have proposed a jet source identification technology. This technology can significantly enhance the scientific discovery capabilities of high-energy collider experiments.

Combining the original high-performance particle flow reconstruction algorithm Arbor and the advanced artificial intelligence technology ParticleNet, the research team developed an efficient injection source identification technology. By capturing the tiny differences between different types of jets, this technology can simultaneously and efficiently distinguish eleven different jets produced by five types of quarks (up, down, strange, charm, and bottom), five types of antiquarks, and gluons. .

This technology can help scientists accurately measure the interactions between different particles, quarks and gluons, and capture some extremely weak Higgs particle decay signals. This technology can improve the accuracy of key physical measurements on future colliders by an order of magnitude, significantly expanding the scientific discovery capabilities of large scientific devices.

The reviewer evaluated it as "the world's top identification performance", "changed the rules of the game", and "created a new vision for accurate measurement in future collider experiments".

The study was titled "Jet-Origin Identification and Its Application at an Electron-Positron Higgs Factory" and was published in "Physical Review Letters" on May 31, 2024.

Paper link: https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.132.221802

Injections can be produced in large quantities in high-energy collision experiments. Accurately identify the origin of the injection, that is, the source of the injection. What kind of quark or gluon is extremely important for discovering and understanding the physical laws behind the collider case. However, the jets produced by different quarks and gluons are very similar in configuration, and it is very difficult to accurately identify the origin of the jets.

Illustration: Event display for a ?+⁢?−→?⁢¯?⁢?→?⁢¯?⁢?⁢? (√?=240 GeV) event simulated and reconstructed using the CEPC baseline detector. (Source: paper)

Identification of jet origins

Background

In order to enhance the scientific discovery capabilities of high-energy collider experiments, researchers have proposed the concept of jet origin identification and divided jets into the following types :

• Five kinds of quarks: a, b, s, u, d
• Five kinds of antiquarks: a¯, b¯, s¯, u¯, d¯
• gluons

Importance

Successful jet source identification is crucial for experimental particle physics at the energy frontier, such as the Large Hadron Collider, as it can effectively reduce the large background in quantum chromodynamics (QCD) processes.

Methods

Researchers used GEANT4-based simulations (called full simulations) to implement the concept of injection origin identification in physics events at the Electron-Positron Higgs Factory, the future The highest priority for the collider project. Researchers have developed software tools such as Arbor and ParticleNet for particle flow event reconstruction and injection source identification.

Performance

The study demonstrates jet origin discrimination performance using an 11-dimensional confusion matrix (M11), which demonstrates the performance of jet "flavor" markers and jet charge measurements.

Illustration: Overall injection origin discrimination performance using 11-dimensional confusion matrix M11. (Source: Paper) Researchers applied jet origin identification to measurements of rare and exotic Higgs boson decays under the CEPC nominal Higgs operational scenario. This scene is expected to be in

Dengan kilauan bersepadu 20 ab^−1 pada GeV, 4 × 10^6 boson Higgs boleh terkumpul. Pereputan jarang berlaku H → ss¯, uu¯ dan dd¯ dan pereputan aliran neutral pengubah rasa (FCNC) H → sb, ds, db dan uc dianalisis (di mana sb bermaksud sb¯ atau s¯b, ds, db dan Begitu juga dengan uc).

Terbitkan had atas untuk tujuh proses ini, antara 10^−3 hingga 10^−4. Dalam Model Standard, nisbah cawangan yang diramalkan untuk proses H → ss¯ ​​​​ ialah 2.3 × 10^−4, dan had atas terbitan adalah bersamaan dengan 3 kali ramalan Model Standard. Nisbah percabangan H → uu¯ dan dd¯ dijangka kurang daripada 10^−6, manakala nisbah percabangan proses FCNC di atas dijangka kurang daripada 10^−7 berdasarkan sumbangan gelung.

Ilustrasi: Kecekapan pelabelan perisa suntikan dan kadar flip cas. (Sumber: Kertas)

Teknologi ini membekalkan para saintis dengan alat baharu yang berkuasa untuk penerokaan saintifik pada masa hadapan, dan juga membuktikan bahawa maklumat kompleks yang terkandung dalam proses penjanaan suntikan boleh difahami dan diproses oleh penggunaan algoritma kecerdasan buatan yang canggih.

Pada masa hadapan, pasukan penyelidik saintifik merancang untuk mengkaji lebih lanjut aplikasi teknologi kecerdasan buatan dalam isu saintifik yang lebih luas dan mendalam.

Pautan kertas: https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.132.221802

Kandungan rujukan:

https://phys.org/news/2024-powered-jetai -identification-technology.html

https://www.ccnta.cn/article/17151.html

The above is the detailed content of Changing the rules of the game for future collider experiments, the Chinese Academy of Sciences and Peking University develop AI-driven injection source identification technology. For more information, please follow other related articles on the PHP Chinese website!