– The Orion cloud complex is the most nearby “mass production factory” of new stars. Since it is closest, it is also easiest to observe and study with telescopes. The part of the complex studied in this work—Orion B—is a very young area. It gives us a possibility to look into what the gas is doing before it is going to collapse and form stars. This is very interesting from the perspective of the star formation process, because it helps us to understand where and how new stars start to form, says Jouni Kainulainen (left), one of the two Chalmers astronomers in the project together with colleague Jan Orkisz, both in the division of Astronomy and Plasma Physics at the Department of Space, Earth and Environment at Chalmers.
With the aim of providing the most detailed analysis yet of the Orion B molecular cloud, the ORION-B team included in its ranks scientists specializing in massive data processing. This enabled them to develop novel methods based on statistical learning and machine learning to study observations of the cloud made at 240000 frequencies of light.
New information from a large mass of data
Based on artificial intelligence algorithms, these tools make it possible to retrieve new information from a large mass of data such as that used in the ORION-B project. This enabled the scientists to uncover a certain number of ‘laws’ governing the Orion molecular cloud.
For instance, they were able to discover the relationships between the light emitted by certain molecules and information that was previously inaccessible, namely, the quantity of hydrogen and of free electrons in the cloud, which they were able to estimate from their calculations – without observing them directly. By analysing all the data available to them, the research team was also able to determine ways of further improving their observations by eliminating a certain amount of unwanted information.
The ORION-B teams now wishes to put their work to a further test, by applying the estimates and recommendations obtained to varying conditions and other star-forming clouds.. Another major theoretical challenge will be to extract information about the speed of molecules, and hence visualize the motion of matter in order to see how it moves within the cloud.
Jouni Kainulainen, what was your and Jan Orkisz’s contributions in this research?
– Jan is a core member of the Orion-B team and has been closely involved in its research for years. In this particular work, he was strongly involved in analyzing the molecular line data that were used. He also contributed in developing the machine learning-based methodology that was exploited in the work. I contribute in the team by helping in the interpretation of some of the analyses and their results.
Were there any results you find particularly interesting for your own research?
– How the gas is distributed in star-forming regions is a difficult, decades old question in astronomy. This work develops a potentially powerful tool to address that question with the help of machine learning. I find this aspect especially intriguing and important: astronomy is a science characterized by large, complex data sets—connection to modern data science is imminent and the benefits of doing so concretize in works like this.
What does this mean for future research about star forming regions?
– The methodology and approach of the paper will hopefully be used more broadly to infer the mass distribution of gas in star forming regions. The molecular lines needed to do the work can be easily observed for example with the ALMA interferometer, which means there is a lot of potential to apply the technique. Doing this could be specifically important in studying clouds more globally in the Milky Way, to understand star-forming regions not only on our “Galactic backyard”, but also in other parts of the galaxy.
Are the new methods applicable to other aspects or subjects in your field?
– Totally. In the essence, the technique uses a set of emission line data to predict the gas mass distribution. This work targeted a young, cold molecular cloud, but just as well one could apply the principle to other kinds of objects in which one is interested in the gas distribution. More broadly speaking, the type of machine learning methodology used in the work is widely applicable—this work demonstrates its use in case of a very specific problem.
The research article mentions that it is “almost impossible to fully understand such stellar nurseries”. What is it like to study near-impossible topics?
– It is almost impossibly interesting! Jokes aside, many systems in nature are so complex that they are very difficult to understand completely—take our brains, for example! But what is amazing is the possibility to comprehend the fundamental basics of how such systems function. Often it turns out that there are, in fact, simple laws and rules that take a long way in understanding those basics. How new stars form in the galaxies is a prime example of this kind of topic, says Jouni Kainulainen, Head of division and assistant professor at the division of Astronomy and Plasma Physics.