Cosmic rayCosmic rays are high-energy particles or clusters of particles (primarily represented by protons or atomic nuclei) that move through space at nearly the speed of light. They originate from the Sun, from outside of the Solar System in our own galaxy, and from distant galaxies. Upon impact with Earth's atmosphere, cosmic rays produce showers of secondary particles, some of which reach the surface, although the bulk is deflected off into space by the magnetosphere or the heliosphere.
Air shower (physics)Air showers are extensive cascades of subatomic particles and ionized nuclei, produced in the atmosphere when a primary cosmic ray enters the atmosphere. When a particle of the cosmic radiation, which could be a proton, a nucleus, an electron, a photon, or (rarely) a positron, interacts with the nucleus of a molecule in the atmosphere, it produces a vast number of secondary particles, which make up the shower.
Ultra-high-energy cosmic rayIn astroparticle physics, an ultra-high-energy cosmic ray (UHECR) is a cosmic ray with an energy greater than 1 EeV (1018 electronvolts, approximately 0.16 joules), far beyond both the rest mass and energies typical of other cosmic ray particles. These particles are extremely rare; between 2004 and 2007, the initial runs of the Pierre Auger Observatory (PAO) detected 27 events with estimated arrival energies above 5.7e19eV, that is, about one such event every four weeks in the 3000 km2 area surveyed by the observatory.
H1 (particle detector)H1 was a particle detector operated at the HERA (Hadron Elektron Ring Anlage) collider at the German national laboratory DESY in Hamburg. The first studies for the H1 experiment were proposed in 1981. The H1 detector began operating together with HERA in 1992 and took data until 2007. It consisted of several different detector components, measured about 12 m × 15 m × 10 m and weighed 2800 tons. It was one of four detectors along the HERA accelerator.
Machine learningMachine learning (ML) is an umbrella term for solving problems for which development of algorithms by human programmers would be cost-prohibitive, and instead the problems are solved by helping machines 'discover' their 'own' algorithms, without needing to be explicitly told what to do by any human-developed algorithms. Recently, generative artificial neural networks have been able to surpass results of many previous approaches.
Cosmic ray spallationCosmic ray spallation, also known as the x-process, is a set of naturally occurring nuclear reactions causing nucleosynthesis; it refers to the formation of chemical elements from the impact of cosmic rays on an object. Cosmic rays are highly energetic charged particles from beyond Earth, ranging from protons, alpha particles, and nuclei of many heavier elements. About 1% of cosmic rays also consist of free electrons. Cosmic rays cause spallation when a ray particle (e.g. a proton) impacts with matter, including other cosmic rays.
Deep learningDeep learning is part of a broader family of machine learning methods, which is based on artificial neural networks with representation learning. The adjective "deep" in deep learning refers to the use of multiple layers in the network. Methods used can be either supervised, semi-supervised or unsupervised.
Particle showerIn particle physics, a shower is a cascade of secondary particles produced as the result of a high-energy particle interacting with dense matter. The incoming particle interacts, producing multiple new particles with lesser energy; each of these then interacts, in the same way, a process that continues until many thousands, millions, or even billions of low-energy particles are produced. These are then stopped in the matter and absorbed. There are two basic types of showers.
Supervised learningSupervised learning (SL) is a paradigm in machine learning where input objects (for example, a vector of predictor variables) and a desired output value (also known as human-labeled supervisory signal) train a model. The training data is processed, building a function that maps new data on expected output values. An optimal scenario will allow for the algorithm to correctly determine output values for unseen instances. This requires the learning algorithm to generalize from the training data to unseen situations in a "reasonable" way (see inductive bias).
Jet (particle physics)A jet is a narrow cone of hadrons and other particles produced by the hadronization of a quark or gluon in a particle physics or heavy ion experiment. Particles carrying a color charge, such as quarks, cannot exist in free form because of quantum chromodynamics (QCD) confinement which only allows for colorless states. When an object containing color charge fragments, each fragment carries away some of the color charge. In order to obey confinement, these fragments create other colored objects around them to form colorless objects.
Deep reinforcement learningDeep reinforcement learning (deep RL) is a subfield of machine learning that combines reinforcement learning (RL) and deep learning. RL considers the problem of a computational agent learning to make decisions by trial and error. Deep RL incorporates deep learning into the solution, allowing agents to make decisions from unstructured input data without manual engineering of the state space. Deep RL algorithms are able to take in very large inputs (e.g.
Wire chamberA wire chamber or multi-wire proportional chamber is a type of proportional counter that detects charged particles and photons and can give positional information on their trajectory, by tracking the trails of gaseous ionization. The multi-wire chamber uses an array of wires at high voltage (anode), which run through a chamber with conductive walls held at ground potential (cathode). Alternatively, the wires may be at ground potential and the cathode held at a high negative voltage; the important thing is that a uniform electric field draws extra electrons or negative ions to the anode wires with little lateral motion.
Q-learningQ-learning is a model-free reinforcement learning algorithm to learn the value of an action in a particular state. It does not require a model of the environment (hence "model-free"), and it can handle problems with stochastic transitions and rewards without requiring adaptations. For any finite Markov decision process (FMDP), Q-learning finds an optimal policy in the sense of maximizing the expected value of the total reward over any and all successive steps, starting from the current state.
Reinforcement learningReinforcement learning (RL) is an area of machine learning concerned with how intelligent agents ought to take actions in an environment in order to maximize the notion of cumulative reward. Reinforcement learning is one of three basic machine learning paradigms, alongside supervised learning and unsupervised learning. Reinforcement learning differs from supervised learning in not needing labelled input/output pairs to be presented, and in not needing sub-optimal actions to be explicitly corrected.
Explainable artificial intelligenceExplainable AI (XAI), also known as Interpretable AI, or Explainable Machine Learning (XML), either refers to an AI system over which it is possible for humans to retain intellectual oversight, or to the methods to achieve this. The main focus is usually on the reasoning behind the decisions or predictions made by the AI which are made more understandable and transparent. XAI counters the "black box" tendency of machine learning, where even the AI's designers cannot explain why it arrived at a specific decision.
Generative adversarial networkA generative adversarial network (GAN) is a class of machine learning framework and a prominent framework for approaching generative AI. The concept was initially developed by Ian Goodfellow and his colleagues in June 2014. In a GAN, two neural networks contest with each other in the form of a zero-sum game, where one agent's gain is another agent's loss. Given a training set, this technique learns to generate new data with the same statistics as the training set.
Particle identificationParticle identification is the process of using information left by a particle passing through a particle detector to identify the type of particle. Particle identification reduces backgrounds and improves measurement resolutions, and is essential to many analyses at particle detectors. Charged particles have been identified using a variety of techniques. All methods rely on a measurement of the momentum in a tracking chamber combined with a measurement of the velocity to determine the charged particle mass, and therefore its identity.
ProtonA proton is a stable subatomic particle, symbol _Proton, H+, or 1H+ with a positive electric charge of +1 e (elementary charge). Its mass is slightly less than that of a neutron and 1,836 times the mass of an electron (the proton-to-electron mass ratio). Protons and neutrons, each with masses of approximately one atomic mass unit, are jointly referred to as "nucleons" (particles present in atomic nuclei). One or more protons are present in the nucleus of every atom.
Alpha particleAlpha particles, also called alpha rays or alpha radiation, consist of two protons and two neutrons bound together into a particle identical to a helium-4 nucleus. They are generally produced in the process of alpha decay, but may also be produced in other ways. Alpha particles are named after the first letter in the Greek alphabet, α. The symbol for the alpha particle is α or α2+. Because they are identical to helium nuclei, they are also sometimes written as He2+ or 42He2+ indicating a helium ion with a +2 charge (missing its two electrons).
Learning rateIn machine learning and statistics, the learning rate is a tuning parameter in an optimization algorithm that determines the step size at each iteration while moving toward a minimum of a loss function. Since it influences to what extent newly acquired information overrides old information, it metaphorically represents the speed at which a machine learning model "learns". In the adaptive control literature, the learning rate is commonly referred to as gain. In setting a learning rate, there is a trade-off between the rate of convergence and overshooting.
