Coughlin part of team that developed alert broker “Babamul” to help parse data from the LSST

The Legacy Survey of Space and Time (LSST), conducted at the Vera C. Rubin Observatory, uses the world’s largest camera to photograph the entire visible night sky over the course of an evening.  The telescope records changes from one image to the next, timed less than a minute apart, sending out a separate alert for each change. With as many as 10,000 changes per minute over an eight-hour period, "that quickly becomes a big data problem," says Associate Professor Michael Coughlin of the Minnesota Institute for Astrophysics. No single astronomer needs all of those alerts, so data scientists have created "brokers" — software that filters the alert stream to surface only the changes relevant to a particular researcher's work. Coughlin is part of a collaboration with Caltech to build a broker called Babamul, which can receive alerts from telescope surveys like Rubin Observatory and the Zwicky Transient Facility (ZTF) to zoom in on those phenomena while they are still happening.

"Our science goals in our hallway at the Minnesota Institute for Astrophysics are all to do with things that go boom: neutron stars, supernovae, tidal disruption of stars. While my graduate students are working on theses about exploding stars or neutron star mergers, if they were to look at those raw alerts, they would have no idea what to do with them. It would be too overwhelming. Babamul is helping them get started in actually making observations in real time."

Coughlin says Babamul is a necessity born from the scale of modern instrumentation. Working with one of the largest telescopes and mirrors in the world means the data rates are unprecedented. 

"The advances in instrumentation are forcing us to make advances in data science to keep pace. We couldn't do LSST without it because the size of the survey represents a huge leap forward in terms of data volume."

The exciting thing about Babamul, Coughlin says, is that it isn't limited to the Rubin Observatory and LSST. When LIGO (Laser Interferometer Gravitational-Wave Observatory) comes back online after upgrades later this year, Babamul will be able to receive gravitational-wave alerts — from say, a potential neutron star merger observable in the southern hemisphere — and follow them with Rubin Observatory to search for the source of those gravitational waves with an entirely different type of instrument. This will allow astronomers to catch incredibly rare phenomena they would almost certainly miss with a traditional telescope. And because the LSST's wide-field-of-view camera covers such an enormous swath of sky, astronomers may not even need to reposition the telescope — they'll just need to know where and when in the observation run to look.

"Because Babamul is in-house," Coughlin says, "we can set the agenda for which specific parameters need to happen to trigger the system." A group studying neutron star mergers, for example, can design the broker stream to suit their specific science goals rather than relying on a one-size-fits-all configuration.

Babamul was built on BOOM, an open-source platform for building telescope data brokers, and is hosted at the Minnesota Supercomputing Institute (MSI) at the University of Minnesota, from which Babamul filters the data and then opens the port to the world.

Senior software developers on the project:

Dr. Sushant Sharma Chaudhary
Antoine Le Calloch

Co-leaders of AppleCiDEr, the machine learning classification system within Babamul:
Felipe Fontinele Nunes
Dr. Argyro Sasli

Student contributors to the project:
Nathan Roe
Anant Gupta
Alexandra Junell
Eric Phillips
Owen Beyer
Caleb Samuel
Benny Border
Maojie Xu
Zheng Lyu
Avi Raghuvanshi
Hailey Markoff
Brayden Rhea
David Kim
Soochun An 
Lingyu Chen
Zewen Yang