By Asro 
In the quest to understand the fundamental nature of our Universe—specifically the mysterious, repulsive force known as dark energy—astronomers have long relied on "standard candles." These are Type Ia supernovae, cosmic explosions of such consistent brightness that they act as reliable distance markers across the vast, dark reaches of space. However, as our instruments have grown more sensitive, these candles have begun to flicker with subtle, unexplained complexities.
Now, a team of researchers led by the Institute of Cosmos Sciences of the University of Barcelona (ICCUB) has unveiled a breakthrough framework that promises to resolve these inconsistencies. Known as CIGaRS, this new methodology merges high-fidelity physical simulations with the predictive power of artificial intelligence to redefine how we measure the expansion of the cosmos. Published in the journal Nature Astronomy, this development arrives just in time to tackle the "data deluge" expected from the upcoming Vera C. Rubin Observatory, potentially increasing the precision of cosmological measurements by a factor of four.
The Pillars of Modern Cosmology: Type Ia Supernovae
To understand why this development is being hailed as a major leap forward, one must first understand the role of Type Ia supernovae. These cataclysmic events occur within binary star systems when a white dwarf—the dense, cooling remnant of a star like our Sun—accretes too much matter from a companion star or merges with another white dwarf. Once the star exceeds a critical mass threshold, it undergoes a thermonuclear runaway, resulting in an explosion of uniform luminosity.
Because these explosions reach a nearly identical peak brightness, they serve as the "standard candles" of the cosmos. By comparing their intrinsic brightness with the light that actually reaches our telescopes, astronomers can calculate their distance from Earth with extraordinary precision. It was these very measurements in the late 1990s that provided the first shocking evidence that the expansion of the Universe is not slowing down due to gravity, but is instead accelerating—a phenomenon attributed to dark energy.
However, the "standard candle" label is a simplification. Over the last two decades, researchers have identified a persistent "nuisance" in their data: the host galaxy effect. A supernova’s appearance is not solely dictated by the physics of the explosion; it is subtly influenced by the environment in which it occurs. Older, more massive galaxies produce supernovae that look systematically different from those in younger, star-forming galaxies. Current correction methods, which rely on simplified approximations, struggle to account for these environmental variations, limiting the accuracy of the cosmological models built upon them.
A Unified Framework: The CIGaRS Architecture
The ICCUB team’s new framework, CIGaRS, represents a paradigm shift in how these variables are processed. Rather than analyzing the supernova, the host galaxy, and the intervening cosmic dust as independent entities, CIGaRS integrates them into a singular, cohesive model.
This unified approach considers the entire lifecycle of the data: the nature of the stellar explosion, the evolutionary history of the host galaxy, the absorption and scattering of light by interstellar dust, and the rate of supernova occurrences throughout cosmic history. By mapping these interdependencies within a single statistical framework, the researchers can capture nuanced relationships that traditional, piece-meal analyses frequently overlook.
"A powerful way of modelling the Universe is to simulate it ab initio in the computer using Bayesian inference," explains Raúl Jiménez, a professor at ICREA-ICCUB and co-author of the study. "This provides a way to vary all possible parameters at the same time to predict what Universe we live in. Furthermore, by having this capacity, one can look into possible ‘unknown unknown’ systematics to understand their effect. The impact of these systematics in our inference is arguably the most important missing ingredient in current approaches to model the Universe."
The Marriage of Artificial Intelligence and Physics
The primary hurdle to building a model of this complexity has always been computational cost. Modeling the interplay of billions of galaxies and millions of supernova events would, under standard circumstances, require an impossible amount of processing power.
The researchers overcame this by employing "simulation-based inference." Instead of calculating every possible outcome through brute-force math, the team trained a deep neural network on a vast library of simulated universes. By feeding the AI thousands of scenarios based on known physical laws, the model learned to recognize the relationship between specific observations—like a galaxy’s color or brightness—and the underlying physical parameters that produced them.
Once the neural network is trained, the process becomes remarkably efficient. It can ingest raw astronomical data and compare it against its "mental map" of simulated universes to determine the most likely physical properties of the real cosmos. This enables the analysis of tens of thousands of supernovae simultaneously, a feat that would be entirely impractical using the standard analytical techniques of the past.
Solving the Redshift Dilemma
One of the most profound implications of the CIGaRS framework is its ability to accurately determine "redshift" using only imaging data. Redshift—the stretching of light caused by the expansion of the Universe—is the primary metric for distance. Traditionally, determining a precise redshift requires spectroscopic follow-up, an expensive and time-consuming process that involves spreading a galaxy’s light into a spectrum to identify chemical signatures.
Because spectroscopic resources are limited, astronomers are typically forced to choose only a small fraction of the supernovae discovered to study in detail. CIGaRS effectively removes this bottleneck. The team demonstrated that their framework could estimate redshifts with accuracy comparable to traditional spectroscopy, but using only photometric data—the simple imaging collected by wide-field surveys.
This capability is a game-changer for the upcoming Vera C. Rubin Observatory. As the observatory begins its decadal survey of the southern sky, it is expected to discover an unprecedented volume of supernovae. Approximately 99% of these objects will be observed only through photometry. Without the CIGaRS approach, the vast majority of these events would be effectively "wasted" for high-precision cosmology.
Preparing for the Rubin Data Deluge
The Vera C. Rubin Observatory is poised to start a new era of "Big Data" in astronomy. As it scans the sky, it will generate a deluge of information that will test the limits of current infrastructure. Konstantin Karchev, the lead author of the study from ICCUB-SISSA Trieste, emphasizes that CIGaRS was engineered specifically for this incoming tidal wave of data.
"Unlike other frameworks, which require analytic simplifications, our no-compromise end-to-end simulation-based inference approach is uniquely capable of extracting the full cosmological and astrophysical information from the Rubin Observatory’s hard-earned data, while avoiding the pitfalls of selection and modelling biases," says Karchev.
By bypassing the need for spectroscopic follow-up for every single event, the CIGaRS framework ensures that the full scientific potential of the Rubin survey can be realized. It allows researchers to treat the entire sky as a laboratory, rather than being limited to the small, biased subsets of supernovae that happened to be selected for spectroscopic analysis.
Beyond Dark Energy: New Insights into Stellar Evolution
While the primary goal of the project is to constrain dark energy and refine our understanding of cosmic expansion, the model offers a secondary, equally compelling benefit: it provides a window into the life cycles of stars.
Because CIGaRS reconstructs how supernova occurrence rates vary alongside the age and chemical composition of stars within different galaxies, it provides scientists with new data regarding the progenitor systems of Type Ia supernovae. By analyzing the "when" and "where" of these explosions with such high granularity, the team is gaining insights into the binary systems that lead to white dwarf explosions, helping to solve long-standing questions about stellar evolution that have remained shrouded in mystery for decades.
The Future of Precision Cosmology
The implications of this research are vast. By integrating physical simulation with the predictive speed of machine learning, the ICCUB team has moved the field of cosmology from a model of "sampling and estimation" to one of "comprehensive reconstruction."
The researchers estimate that their approach could improve cosmological constraints by up to a factor of four compared to traditional, spectroscopy-dependent methods. This is not merely an incremental gain; it is a fundamental shift in how we "see" the history and future of the Universe.
As the Vera C. Rubin Observatory prepares for its first light, the CIGaRS framework stands as a critical piece of the puzzle. It ensures that when the data begins to flow, the astronomical community will be ready not just to store it, but to truly understand it. In the coming decade, as we look deeper into the accelerating expansion of our Universe, the work of the ICCUB team will likely serve as the bedrock upon which our next major discoveries in physics are built. The dark energy that drives our Universe remains elusive, but thanks to CIGaRS, our ability to measure its effects has never been clearer.
Ali Ikhwan 
Reynand Wu 
Ammar Sabilarrohman 