From Carnegie Mellon University: “Hey Siri:: How Much Does This Galaxy Cluster Weigh?”

From Carnegie Mellon University

July 19, 2022
Amy Pavlak Laird

Jocelyn Duffy
Mellon College of Science

It’s been nearly a century since astronomer Fritz Zwicky first calculated the mass of the Coma Cluster, a dense collection of almost 1,000 galaxies located in the nearby universe. But estimating the mass of something so huge and dense, not to mention 320 million light-years away, has its share of problems — then and now. Zwicky’s initial measurements, and the many made since, are plagued by sources of error that bias the mass higher or lower.

Now, using tools from machine learning, a team led by Carnegie Mellon University physicists has developed a deep-learning method that accurately estimates the mass of the Coma Cluster and effectively mitigates the sources of error.

“People have made mass estimates of the Coma Cluster for many, many years. But by showing that our machine-learning methods are consistent with these previous mass estimates, we are building trust in these new, very powerful methods that are hot in the field of cosmology right now,” said Matthew Ho, a fifth-year graduate student in the Department of Physics’ McWilliams Center for Cosmology and a member of Carnegie Mellon’s NSF AI Planning Institute for Physics of the Future.

Machine-learning methods are used successfully in a variety of fields to find patterns in complex data, but they have only gained a foothold in cosmology research in the last decade. For some researchers in the field, these methods come with a major concern: Since it is difficult to understand the inner workings of a complex machine-learning model, can they be trusted to do what they are designed to do? Ho and his colleagues set out to address these reservations with their latest research.

To calculate the mass of the Coma Cluster, Zwicky and others used a dynamical mass measurement, in which they studied the motion or velocity of objects orbiting in and around the cluster and then used their understanding of gravity to infer the cluster’s mass. But this measurement is susceptible to a variety of errors. Galaxy clusters exist as nodes in a huge web of matter distributed throughout the universe, and they are constantly colliding and merging with each other, which distorts the velocity profile of the constituent galaxies.

And because astronomers are observing the cluster from a great distance, there are a lot of other things in between that can look and act like they are part of the galaxy cluster, which can bias the mass measurement. Recent research has made progress toward quantifying and accounting for the effect of these errors, but machine-learning-based methods offer an innovative data-driven approach, according to Ho.

“Our deep-learning method learns from real data what are useful measurements and what are not,” Ho said, adding that their method eliminates errors from interloping galaxies (selection effects) and accounts for various galaxy shapes (physical effects). “The usage of these data-driven methods makes our predictions better and automated.”

“One of the major shortcomings with standard machine learning approaches is that they usually yield results without any uncertainties,” added Associate Professor of Physics Hy Trac, Ho’s adviser. “Our method includes robust Bayesian statistics, which allow us to quantify the uncertainty in our results.”

Ho and his colleagues developed their novel method by customizing a well-known machine-learning tool called a convolutional neural network, which is a type of deep-learning algorithm used in image recognition. The researchers trained their model by feeding it data from cosmological simulations of the universe. The model learned by looking at the observable characteristics of thousands of galaxy clusters, whose mass is already known. After in-depth analysis of the model’s handling of the simulation data, Ho applied it to a real system — the Coma Cluster — whose true mass is not known. Ho’s method calculated a mass estimate that is consistent with most of the mass estimates made since the 1980s. This marks the first time this specific machine-learning methodology has been applied to an observational system.

“To build reliability of machine-learning models, it’s important to validate the model’s predictions on well-studied systems, like Coma,” Ho said. “We are currently undertaking a more rigorous, extensive check of our method. The promising results are a strong step toward applying our method on new, unstudied data.”

Models such as these are going to be critical moving forward, especially when large-scale spectroscopic surveys, such as the Dark Energy Spectroscopic Instrument, the Vera C. Rubin Observatory and Euclid, start releasing the vast amounts of data they are collecting of the sky.

“Soon we’re going to have a petabyte-scale data flow,” Ho explained. “That’s huge. It’s impossible for humans to parse that by hand. As we work on building models that can be robust estimators of things like mass while mitigating sources of error, another important aspect is that they need to be computationally efficient if we’re going to process this huge data flow from these new surveys. And that is exactly what we are trying to address — using machine learning to improve our analyses and make them faster.”

This work is supported by NSF AI Institute: Physics of the Future, NSF PHY-2020295, and the McWilliams-PSC Seed Grant Program. The computing resources necessary to complete this analysis were provided by the Pittsburgh Supercomputing Center.

Pittsburgh Supercomputer Center


The CosmoSim database used in this paper is a service by the Leibniz-Institute for Astrophysics Potsdam (AIP).

The study’s authors include: Trac; Michelle Ntampaka, who graduated from CMU with a doctorate in physics in 2017 and is now deputy head of Data Science at the Space Telescope Science Institute; Markus Michael Rau, a McWilliams postdoctoral fellow who is now a postdoctoral fellow at Argonne National Lab; Minghan Chen, who graduated with a bachelor’s degree in physics in 2018, and is a Ph.D. student at the University of California, Santa Barbara.; Alexa Lansberry, who graduated with a bachelor’s degree in physics in 2020; and Faith Ruehle, who graduated with a bachelor’s degree in physics in 2021.

Science paper:
Nature Astronomy

See the full article here .


Please help promote STEM in your local schools.

Stem Education Coalition

Carnegie Mellon University is a global research university with more than 12,000 students, 95,000 alumni, and 5,000 faculty and staff.
CMU has been a birthplace of innovation since its founding in 1900.
Today, we are a global leader bringing groundbreaking ideas to market and creating successful startup businesses.
Our award-winning faculty members are renowned for working closely with students to solve major scientific, technological and societal challenges. We put a strong emphasis on creating things—from art to robots. Our students are recruited by some of the world’s most innovative companies.

We have campuses in Pittsburgh, Qatar and Silicon Valley, and degree-granting programs around the world, including Africa, Asia, Australia, Europe and Latin America.

The university was established by Andrew Carnegie as the Carnegie Technical Schools, the university became the Carnegie Institute of Technology in 1912 and began granting four-year degrees. In 1967, the Carnegie Institute of Technology merged with the Mellon Institute of Industrial Research, formerly a part of the University of Pittsburgh. Since then, the university has operated as a single institution.

The university has seven colleges and independent schools, including the College of Engineering, College of Fine Arts, Dietrich College of Humanities and Social Sciences, Mellon College of Science, Tepper School of Business, Heinz College of Information Systems and Public Policy, and the School of Computer Science. The university has its main campus located 3 miles (5 km) from Downtown Pittsburgh, and the university also has over a dozen degree-granting locations in six continents, including degree-granting campuses in Qatar and Silicon Valley.

Past and present faculty and alumni include 20 Nobel Prize laureates, 13 Turing Award winners, 23 Members of the American Academy of Arts and Sciences, 22 Fellows of the American Association for the Advancement of Science , 79 Members of the National Academies, 124 Emmy Award winners, 47 Tony Award laureates, and 10 Academy Award winners. Carnegie Mellon enrolls 14,799 students from 117 countries and employs 1,400 faculty members.

Carnegie Mellon University is classified among “R1: Doctoral Universities – Very High Research Activity”. For the 2006 fiscal year, the university spent $315 million on research. The primary recipients of this funding were the School of Computer Science ($100.3 million), the Software Engineering Institute ($71.7 million), the College of Engineering ($48.5 million), and the Mellon College of Science ($47.7 million). The research money comes largely from federal sources, with a federal investment of $277.6 million. The federal agencies that invest the most money are the National Science Foundation and the Department of Defense, which contribute 26% and 23.4% of the total university research budget respectively.

The recognition of Carnegie Mellon as one of the best research facilities in the nation has a long history—as early as the 1987 Federal budget Carnegie Mellon University was ranked as third in the amount of research dollars with $41.5 million, with only Massachusetts Institute of Technology and Johns Hopkins University receiving more research funds from the Department of Defense.

The Pittsburgh Supercomputing Center is a joint effort between Carnegie Mellon, University of Pittsburgh, and Westinghouse Electric Company. Pittsburgh Supercomputing Center was founded in 1986 by its two scientific directors, Dr. Ralph Roskies of the University of Pittsburgh and Dr. Michael Levine of Carnegie Mellon. Pittsburgh Supercomputing Center is a leading partner in the TeraGrid, the National Science Foundation’s cyberinfrastructure program.

Scarab lunar rover is being developed by the RI.

The Robotics Institute (RI) is a division of the School of Computer Science and considered to be one of the leading centers of robotics research in the world. The Field Robotics Center (FRC) has developed a number of significant robots, including Sandstorm and H1ghlander, which finished second and third in the DARPA Grand Challenge, and Boss, which won the DARPA Urban Challenge. The Robotics Institute has partnered with a spinoff company, Astrobotic Technology Inc., to land a CMU robot on the moon by 2016 in pursuit of the Google Lunar XPrize. The robot, known as Andy, is designed to explore lunar pits, which might include entrances to caves. The RI is primarily sited at Carnegie Mellon’s main campus in Newell-Simon hall.

The Software Engineering Institute (SEI) is a federally funded research and development center sponsored by the U.S. Department of Defense and operated by Carnegie Mellon, with offices in Pittsburgh, Pennsylvania, USA; Arlington, Virginia, and Frankfurt, Germany. The SEI publishes books on software engineering for industry, government and military applications and practices. The organization is known for its Capability Maturity Model (CMM) and Capability Maturity Model Integration (CMMI), which identify essential elements of effective system and software engineering processes and can be used to rate the level of an organization’s capability for producing quality systems. The SEI is also the home of CERT/CC, the federally funded computer security organization. The CERT Program’s primary goals are to ensure that appropriate technology and systems management practices are used to resist attacks on networked systems and to limit damage and ensure continuity of critical services subsequent to attacks, accidents, or failures.

The Human–Computer Interaction Institute (HCII) is a division of the School of Computer Science and is considered one of the leading centers of human–computer interaction research, integrating computer science, design, social science, and learning science. Such interdisciplinary collaboration is the hallmark of research done throughout the university.

The Language Technologies Institute (LTI) is another unit of the School of Computer Science and is famous for being one of the leading research centers in the area of language technologies. The primary research focus of the institute is on machine translation, speech recognition, speech synthesis, information retrieval, parsing and information extraction. Until 1996, the institute existed as the Center for Machine Translation that was established in 1986. From 1996 onwards, it started awarding graduate degrees and the name was changed to Language Technologies Institute.

Carnegie Mellon is also home to the Carnegie School of management and economics. This intellectual school grew out of the Tepper School of Business in the 1950s and 1960s and focused on the intersection of behavioralism and management. Several management theories, most notably bounded rationality and the behavioral theory of the firm, were established by Carnegie School management scientists and economists.

Carnegie Mellon also develops cross-disciplinary and university-wide institutes and initiatives to take advantage of strengths in various colleges and departments and develop solutions in critical social and technical problems. To date, these have included the Cylab Security and Privacy Institute, the Wilton E. Scott Institute for Energy Innovation, the Neuroscience Institute (formerly known as BrainHub), the Simon Initiative, and the Disruptive Healthcare Technology Institute.

Carnegie Mellon has made a concerted effort to attract corporate research labs, offices, and partnerships to the Pittsburgh campus. Apple Inc., Intel, Google, Microsoft, Disney, Facebook, IBM, General Motors, Bombardier Inc., Yahoo!, Uber, Tata Consultancy Services, Ansys, Boeing, Robert Bosch GmbH, and the Rand Corporation have established a presence on or near campus. In collaboration with Intel, Carnegie Mellon has pioneered research into claytronics.