Scientists reveal the limits of machine learning for hydrogen models
All of you know about the hydrogen gas which is a clear option to methane, also called as natural gas. It's the most abundant chemical element, estimated to contribute 75% of the mass of the universe. On the earth, vast numbers of hydrogen atoms are contained in water, plants, animals and, of course, humans. But while it’s present in nearly all molecules in living things, it’s very scarce as a gas smaller than one part per million by volume.
Hydrogen may be formed from a many types of resources, such as natural gas, nuclear power, biogas and renewable power like solar and wind. The challenge is harnessing hydrogen as a gas on a large scale to fuel our homes and businesses.
We know that, the occurrence about the hydrogen on the Earth. Hydrogen is actually occurs as a gas, But if it is under maximum temperatures and pressures, the situation that occurs within many planets, like Jupiter- hydrogen goes through a series of state transitions and takes on the properties of a liquid metal. One of the most important metallic characteristic it takes on is becoming an electrical conductor.
In a latest study in the Nature journal's "Matters Arising," scientists at the University of Rochester Laboratory for Laser Energetics (LLE), containing lead author Valentin Karasiev, an LLE staff scientist; graduate student Josh Hinz; and Suxing Hu, an associate professor of mechanical engineering and a distinguished scientist at the LLE, observed to a 2020 Nature paper that used machine learning techniques to study the liquid-liquid phase transitions of dense hydrogen from an insulating liquid to a liquid metal.
In their study, Karasiev and his colleagues outline how these machine learning techniques created fault in results in understanding hydrogen's phase transitions. Their study has most necessary implications in building more correct computer models to research hydrogen,that may be lead to a better understanding of the interiors of planets and stars and the physical properties of processes such as nuclear fusion. When creating the equation of state of hydrogen the equation which defines the state of hydrogen under several physical conditions, it is necessary to characterize the transition into the metallic hydrogen phase- Is it an Sharp transition or a smooth transition?
The Hu-says that- "This physics character of first-order phase transition may have profound implications in describing what giant planets interior structures look like, such as de-mixing of hydrogen and helium in Jupiter".
In the 2020 Nature paper, scientists used machine learning and observed the transition of hydrogen to the metallic hydrogen phase was smooth. Karasiev and his colleagues, however, performed large-scale quantum simulations using other fundamental density-functional theory and observed that hydrogen's transition is not smooth, but is instead more abrupt. This is consistent with other previous data gathered without machine learning.
The Karasiev says that- "Our work demonstrated that machine learning can fool researchers if they are not worry when using machine learning to research phase-transition boundaries. This is an necessary method in building better models to outline how hydrogen can become metallic hydrogen".
