Computational Astrophysics

In the past humans have used their own intuition and knowledge to understand their environments.  To an extent, this approach is beneficial to the betterment of science. But there is a limit to how much information one human can hold and how quickly they can use that information to solve problems– this is where computers come in. Ever since the birth of modern computers, they have been used to solve complex problems that no one human can solve and with the speed that no one human can think in. One of the most interesting applications of this unprecedented storage and speed that computers have is in the field of astronomy. With there being 100 billion galaxies and each of them having around 100 billion stars, absolutely no humans can analyze all of these, and not only would a computer be helpful, but it is also necessary for modern astrophysics. A form of machine learning and unsupervised learning is what is used in astronomy due to its ability to find patterns without being told the labels or outputs in a dataset explicitly. Methods include clustering, classification, dimensionality reduction, and association. Moreover, computational simulations are popular in astronomy as well. They give scientists a way to understand celestial events. 

With these algorithms, computers can do many things from calculating certain aspects of a celestial body or event, it can predict certain features of a celestial object, or it can just classify it with simple linear classification or clustering. Other than machine learning, computers are also used to simulate celestial events, such as supernovae, star births, planet formations, and more. Since we know the laws and rules that these celestial bodies live by, we can create simulations to see how these bodies would act under different circumstances. For example, computer simulations gave us images of black holes before the first true image five years ago. The properties and features of the black hole are put into a simulation with the rules and laws that it follows and a simulation.   Although these simulations are not entirely accurate when you compare them to the actual image of a black hole, they still accomplished their task of simulating the behavior of a black hole and how a black hole looks to a certain extent.