As a child, I had the opportunity to use the two-meter Faulkes telescope in Siding Spring, Australia. I remember seeing the telescope moving through a live webcam was a very exciting experience. I would enter air coordinates and over the course of a minute the telescope would slowly tilt toward the object. I couldn’t wait to become an astronomer and spend countless nights in the actual telescope locations – Hawaii, Chile, Spain, these are all locations with powerful telescopes I ever wanted to use.
But I didn’t know that the era of manual observations was coming to an end.
My first professional research experience came when I was hired as an intern at Las Cumbres Observatory, headquartered in Santa Barbara, California, my hometown. Although it’s called an observatory, it doesn’t actually contain a telescope. Las Cumbres includes more than 30 instruments around the world, with the mission of always having a telescope in the dark, ready to make observations in the blink of an eye.
Their telescope system is very streamlined and is a great example of how astronomical observations work today. The telescopes are all robotically controlled, requiring almost no human intervention, with the exception of selected engineers who are responsible for maintaining the instruments. Scheduling is also robotically controlled, and while astronomers may request observations during certain nights, the observations are ultimately booked and monitored by an automated system programmed to select the best targets depending on instrument availability and weather conditions at all locations. ; an analysis performed in seconds.
Indeed, it is becoming increasingly rare for astronomers to visit sites and make actual observations, as manual operation of telescopes, when necessary, is often done remotely from the comfort of one’s home; and every collaboration is easy via zoom. Gone are the days of Edwin Hubble, who would actually live on the mountain for extended periods, observing night after night and collecting records of data one by one.
While I find it a little sad that such days are behind us, there is a silver lining to it all. With technological advances in both telescope instrumentation and software capabilities, it became necessary to rethink and streamline the observation process. We had no other choice, because humans are of course prone to mistakes, but they also need to sleep, eat and socialize, while a machine can work non-stop for decades as long as there is electricity. As an astronomer, I value science above all else and would gladly prefer an increase in quality data over the ability to live on a mountain and make manual observations, however mystical and dreamy that may be.
Machine learning has been used since the 20th century as a tool to optimize manual procedures. In fact, my colleagues used to joke that the USPS mail system has been using machine learning to help sort the mail since the 1960s, while astronomers have recently come to appreciate its application in our field because of the big data transition we are in. The James Webb telescope, for example, will produce more data than can ever be visually inspected. There are images today that no one in this world will ever see, and with the advent of new telescopes in the coming years, more images will fall into the category of data only inspected by emotionless machines.
In short, machine learning is the use of differential calculus to identify optimal patterns in high-dimensional data. For example, a 50 by 50 pixel image can be displayed in 2500 dimensional space. But what are these “optimal patterns” that the machine identifies? Unfortunately there is no answer. Really, machine learning is often seen as a dark art. Even if we could visualize the connections of my machine learning engine, it would have been pointless, because we would have had little understanding of it in the end. The correlations the machine has found through countless iterations are just too complicated for our minds to comprehend. It’s really a black box – data is coming in, we don’t know what connections the machine learned during training, but good results are coming out and we’re happy.
As astronomers, we use machine learning for object recognition, signal prediction and even as a tool to manage our instruments. It would take me my whole life to inspect two million astronomical objects, but I have a machine learning algorithm that did it for me in less than 30 minutes. These developments have led to the creation of broker systems that take telescope data, apply machine learning to distinguish particular objects, and then forward the information to scientific teams interested in the specific phenomena.
Much data will go undetected for the next century, even with the help of our machine learning programs; but I suppose that’s a beautiful thing – anyone can do astronomy simply by downloading public data from their computer. We need all the help we can get because while machine learning has tremendous utility, in its current state it still cannot be compared to the eyes and brains we are gifted with, of which there simply aren’t enough.
Daniel Godines is a PhD student in astronomy at New Mexico State University. He can be reached at [email protected]