Trieste, 6 May 2026 — In the quest to decipher the grand narrative of our Universe’s expansion, Type Ia supernovae have long served as indispensable tools for astronomers. These brilliant stellar explosions act as cosmic lighthouses, allowing scientists to gauge vast intergalactic distances by comparing their intrinsic brightness to their observed luminosity. However, extracting precise cosmological insights from their light curves is fraught with complexity, complicated by a tapestry of intrinsic and extrinsic influences that modify the signal before it reaches Earth. Addressing this formidable challenge, researchers Konstantin Karchev and Roberto Trotta of SISSA, alongside Raúl Jiménez of the University of Barcelona, have pioneered a groundbreaking approach that harnesses artificial intelligence to extract unprecedented detail solely from supernova brightness data.

Type Ia supernovae are prized in cosmology primarily because of their reputation as “standardisable candles.” This designation implies that their inherent luminosity, while not perfectly uniform, can be calibrated through empirical relationships, enabling astronomers to measure cosmic distances with remarkable accuracy. Nevertheless, the nuance lies in the fact that their apparent brightness is influenced not only by the physics inherent to the explosion but also by the evolutionary history and environment of the progenitor star. Factors such as stellar age, metallicity, and the interplay with interstellar dust within the host galaxy convolute the light we observe, presenting an interpretative labyrinth for astrophysicists.

Traditionally, spectroscopic analysis has been the gold standard for disentangling these layers of complexity. Spectroscopy offers vital clues by decomposing the supernova light into its constituent wavelengths, revealing fingerprints of the explosion’s chemistry and surrounding environment. Yet, acquiring high-quality, homogeneous spectral data across large supernova samples is logistically and financially prohibitive, especially as upcoming surveys promise to deliver millions of new detections. In this landscape, photometric data—which records brightness over time and across broad filter bands—stands as a more attainable but less informative alternative, requiring innovative data analysis methods to unlock its full potential.

A historical crutch in the field has been the so-called “mass step” correction. Observations have shown that Type Ia supernovae in galaxies exceeding a certain stellar mass threshold (~10 billion solar masses) exhibit systematically different luminosities compared to those in less massive hosts. As a pragmatic if imperfect solution, astronomers have applied a step correction based on galaxy mass, serving as a proxy for multiple underlying physical factors influencing supernova brightness. While this technique has marginally improved standardisation, it remains a coarse and indirect correction that homogenises a diversity of stellar and galactic conditions into a single binary parameter.

Enter CIGaRS — Combined Inference and Galaxy-Related Standardisation — an ingenious method that revolutionises the analysis of Type Ia supernovae photometric data. Developed using state-of-the-art neural network architectures, CIGaRS synthesizes multiple astrophysical processes into a unified probabilistic model. This method simultaneously integrates galaxy evolutionary models, dust attenuation physics, supernova delay-time distributions, and the intrinsic properties of the explosions themselves. Unlike previous approaches that treat galaxy mass, dust effects, and progenitor characteristics as separate correction steps, CIGaRS holistically decodes the observed luminosity variations, enabling a simultaneous and self-consistent inference of underlying causes.

Testing their method rigorously, the research team first constructed an extensive simulated catalogue emulating real-world supernova datasets, incorporating 1,578 carefully selected supernovae to resemble contemporary samples. They then extrapolated to a vastly larger dataset of approximately 16,000 objects, mirroring the scale of data anticipated from the Vera Rubin Observatory’s Legacy Survey of Space and Time (LSST) over just a single month. The results were nothing short of remarkable. By leveraging only photometry, CIGaRS effectively inferred critical properties that were previously accessible only through detailed spectroscopic campaigns.

Crucially, CIGaRS not only recovers cosmological parameters—such as those dictating the Universe’s expansion rate—but also untangles the delay-time distribution that governs how long after a star’s birth it detonates as a Type Ia supernova. Moreover, it differentiates the subtle imprints left by progenitor stellar age and chemical composition on the luminosity distribution. The model delineates that chemical composition tends to manifest effects mimicking the classic “mass step,” with luminosity adjustments correlated to progenitor metallicity, whereas age impacts introduce smoother gradients across observed brightnesses. This nuanced understanding fundamentally advances how astronomers interpret subtle variances in supernova magnitudes observed within diverse galactic environments.

One of the central challenges addressed by CIGaRS lies in deconvolving these small but critical effects from dominant sources of variability like light colour and dust extinction. Standard analytical techniques often stumble at this task due to overlapping signatures and limited data fidelity. By contrast, the AI-based approach excels at recognizing complex, nonlinear patterns across the multi-dimensional photometric parameter space, effectively peeling back layers that previously obscured key astrophysical insights.

The transformative implications for cosmology are profound. Traditionally, only a small fraction of supernovae detected photometrically are follow-up with spectroscopy—usually around one percent—substantially limiting the precision of cosmological measurements. CIGaRS empowers astronomers to harness the overwhelming majority of photometric-only supernova observations effectively, enhancing the precision of parameter estimation by approximately a factor of four. This leap in precision could dramatically sharpen constraints on dark energy models, the Hubble constant, and other pivotal cosmological metrics, accelerating our understanding of the Universe’s past and future dynamics.

The imminent influx of supernova data from LSST and other next-generation surveys crystallizes the urgency for methods like CIGaRS. As Roberto Trotta, theoretical physics professor at SISSA, emphasizes, future observational datasets will be too vast and complex for classical analytic techniques. Innovative computational tools powered by machine learning are no longer optional enhancements but essential instruments for mining transformative science from the impending data deluge.

By deconstructing the interplay between intrinsic supernova physics and extrinsic environmental influences, CIGaRS marks a paradigm shift in how we calibrate cosmic distance indicators. This integrated framework heralds a future where photometric supernova surveys, far less resource-intensive than their spectroscopic counterparts, can deliver cosmological insights with unprecedented clarity and depth. As the observational capabilities of humanity’s telescopes reach new frontiers, so too must our analytic techniques evolve—melding astrophysical theory with cutting-edge artificial intelligence to illuminate the expanding Universe in ever finer detail.

This pioneering study not only refines cosmological measurements but sets a precedent for exploiting vast, heterogenous astronomical datasets using simulation-based inference and neural networks. The era of “data-rich, insight-poor” astrophysics is ending; in its place comes a bold vision of comprehensive understanding propelled by smart algorithms capable of translating subtle signals into fundamental knowledge. As this methodology matures, it promises to unlock new physics and deepen our grasp of the cosmic story written in the light of dying stars.

Subject of Research: Not applicable

Article Title: CIGaRS I: combined simulation-based inference from type Ia supernovae and host photometry

News Publication Date: 6-May-2026

Web References: https://doi.org/10.1038/s41550-026-02842-5

Keywords

Type Ia supernovae, photometry, cosmology, artificial intelligence, neural networks, cosmic expansion, standardisable candles, supernova progenitor, galaxy evolution, simulation-based inference, Vera Rubin Observatory, Legacy Survey of Space and Time (LSST), dust extinction, stellar metallicity, supernova delay-time distribution

Since its landmark achievement in 2015, when LIGO accomplished the first direct detection of gravitational waves—a scientific milestone that vindicated Einstein’s century-old predictions—the field of gravitational-wave astronomy has burgeoned. This pivotal discovery culminated in the awarding of the Nobel Prize in Physics in 2017 to three of LIGO’s lead scientists. In the ensuing years, enhancements to LIGO’s ongoing experiments have allowed the observatory to register approximately one black hole merger every three days, vastly expanding our understanding of these enigmatic cosmic entities. Alongside its international collaborators—the Virgo gravitational-wave detector situated in Italy and KAGRA in Japan—LIGO has unearthed hundreds of candidates for black hole mergers, revealing a wealth of data that were previously inaccessible.

The research community at LIGO is steadfast in its commitment to augmenting the observatory’s capabilities, particularly in identifying a wider array of black hole mergers. One specific area of interest pertains to the potential discovery of more massive mergers that may inhabit a theorized intermediate-mass range bridging the gap between stellar-mass black holes and the supermassive black holes that reside at the centers of galaxies. By enhancing LIGO’s sensitivity, researchers aim to detect black holes with more eccentric orbits and capture merging events at earlier stages of their coalescence when the cosmic bodies spiral closer together.

To facilitate this ambitious goal, a collaborative effort between Caltech, the Gran Sasso Science Institute in Italy, and Google DeepMind has initiated the development of a cutting-edge AI methodology termed Deep Loop Shaping. This innovative approach focuses on dramatically improving the suppression of unwanted noise within LIGO’s detectors. In scientific parlance, “noise” encompasses various disruptive background disturbances that can compromise the integrity of data collection. While such noise can manifest as literal sound waves, it typically refers to subtle fluctuations in the highly sensitive mirrors crucial to LIGO’s functionality. Minimizing these disturbances is essential for accurately capturing the telltale signals of gravitational waves.

In a recent publication in the journal Science, it was reported that the AI algorithm designed through this collaboration successfully quieted the movements of LIGO’s mirrors by a factor of 30 to 100 times greater than traditional noise-reduction technologies could achieve. This is a pioneering achievement, as it establishes a new standard in the quest for precision measurement in gravitational-wave detection. Co-author and leading researcher Rana Adhikari, a professor of physics at Caltech, encapsulated the groundbreaking nature of this technology by stating that it enhances LIGO’s ability to identify more substantial black holes and beyond, potentially paving the way for the next generation of even more sophisticated gravitational-wave observatories.

The implications of this research extend far beyond astrophysics alone. The principles underlying Deep Loop Shaping have the potential to reverberate throughout various engineering disciplines, especially those predicated upon control systems. As study co-authors Brendan Tracey and Jonas Buchli from Google DeepMind noted, this methodology could find applications in diverse fields including aerospace, robotics, and structural engineering, where vibration suppression and noise cancellation are critical to success.

LIGO’s impressive structure consists of two “L” shaped facilities where each arm houses a vacuum tube engineered to facilitate advanced laser technology. These tubes, measuring approximately 4 kilometers in length, host powerful lasers that reflect back and forth utilizing colossal 40-kilogram mirrors positioned at either end. As gravitational waves traverse Earth from astronomical events, they distort space-time in a manner that leads to minute changes in the lengths of the arms, which LIGO’s laser system is specifically designed to detect. However, to achieve the extraordinary precision required for such measurements, engineers must strive to mitigate any background noise that could interfere with the delicate operation.

The study delineated how oceanic activity stands as one of the primary disruptors of LIGO’s mirror stability, causing vibrations transmitted through the ground that can sway the mirrors even when the facilities are situated far from coastal areas. Co-author Christopher Wipf offered a colorful analogy, likening noise cancellation in LIGO to noise-canceling headphones that use external microphones to detect and counteract unwanted environmental sounds. The controls in place at LIGO operate on a feedback system, akin to managing vibrations on a waterbed—a balancing act that involves compensating for disturbances while simultaneously avoiding the introduction of new, unintended vibrations.

The challenge for LIGO engineers lies in addressing this “hiss” of self-induced noise within the control system itself. Traditional feedback controllers operate effectively by sensing seismic disturbances and counteracting them, but in the process, they can inadvertently generate higher-frequency noise that further complicates data collection. To better manage these complexities, the collaboration initiated efforts to enhance the control system using AI methodologies.

The journey began approximately four years ago when Jan Harms, a dedicated researcher previously affiliated with Caltech, reached out to Google DeepMind’s experts to explore artificial intelligence as a solution for better managing the vibrations affecting LIGO’s mirrors. The team subsequently engaged in extensive trials of various AI techniques, ultimately focusing on reinforcement learning—an approach enabling the algorithm to learn control strategies through repeated simulations. By generating numerous simulations of LIGO to optimize performance, the AI ultimately demonstrated a remarkable capacity for noise suppression, contributing to the observatory’s overarching mission.

Richard Murray, a professor of Control and Dynamical Systems at Caltech, underscored the dual significance of this research. It not only represents a technical advancement in gravitational-wave detection but also showcases AI’s capacity to enhance control systems across an array of complex applications. This revelation encourages a new generation of scientists and engineers to engage with LIGO, fueling innovation at the cutting edge of modern technology and measurement science.

Although initial trials using the new AI method were limited to just an hour, the research team is poised to conduct longer and more thorough tests in the near future. As they work towards deploying this innovative solution on several LIGO systems, the potential that has been unlocked introduces exciting possibilities for the future of gravitational-wave detection. By fundamentally altering how we approach the challenges associated with ground-based detection methods, the implications of this research branch into multiple domains of science and technology.

As LIGO continues to unravel the mysteries of the universe, this new AI methodology represents a paradigm shift, enabling researchers to navigate complex variables in gravitational-wave detection with enhanced precision. The journey is only beginning, and as we stand at the precipice of a new era in astrophysics, the promise of AI could redefine our capability to probe the depths of space and time like never before.

Subject of Research: Enhancing LIGO’s detection capabilities using AI
Article Title: Improving cosmological reach of a gravitational wave observatory using Deep Loop Shaping
News Publication Date: 4-Sep-2025
Web References: DOI: 10.1126/science.adw1291
References: N/A
Image Credits: Caltech/MIT/LIGO Lab

Keywords

Gravitational waves, LIGO, AI, Deep Loop Shaping, astrophysics, black holes, control systems, noise cancellation, vibration suppression, space-time detection, scientific innovation, advanced measurement techniques.

Astrophysicists face the monumental task of understanding how galaxies form and evolve. While phenomena like supernovae are known to play a crucial role in this process, direct observations of such events in the cosmos can be limited. Consequently, researchers rely heavily on numerical simulations that draw from vast datasets sourced from telescopes and various instruments that analyze interstellar space. These simulations must account for multiple factors, including gravitational forces, hydrodynamics, and the intricate nature of astrophysical thermo-chemistry.

One of the most pressing challenges in conducting galaxy evolution simulations lies in achieving high temporal resolution. Researchers aim to generate 3D snapshots of a galaxy’s evolution with very short time intervals, ensuring that critical events, like the expansion of supernova shells, are accurately captured. However, current supercomputing techniques often miss out on these rapid phenomena due to their limited temporal scopes—standard simulations take years to complete, even for relatively small galaxies.

Tackling the notorious “timestep bottleneck” was the key objective of Hirashima and his team’s research. By integrating machine learning into their simulation framework, they managed to replicate the results of a previously established model of a dwarf galaxy while optimizing processing time. Hirashima stated, “When we use our AI model, the simulation is about four times faster than conventional numerical simulations.” This advancement has the potential to reduce computation time from several months to just a few weeks, marking a significant leap forward in the field.

The AI-assisted simulation framework is powered by a neural network that has been trained on an extensive dataset comprising 300 simulations. These simulations focused on isolated supernova events occurring within a molecular cloud with a mass equivalent to about one million solar masses. After the model was trained, it became adept at predicting critical parameters such as density, temperature, and the 3D velocities of gas within 100,000 years post-explosion.

The results produced by the AI-driven model were nothing short of remarkable. Not only did the new framework achieve similar structural outcomes and star formation histories as traditional simulation techniques, but it did so in a fraction of the time. By reducing the computation load, this advanced methodology opens the door to conducting high-resolution simulations of larger galaxies, such as the Milky Way, which have often been constrained by the limitations of standard supercomputers.

Looking forward, Hirashima envisions that this transformative AI-assisted framework could lead to star-by-star simulations of massive galaxies with the intent of unraveling the mysteries surrounding the origin of our solar system, as well as identifying the essential elements required for life. Currently, the research team is leveraging this new model to simulate a Milky Way-sized galaxy, which could yield unprecedented insights into the conditions that gave rise to life on Earth.

The potential ramifications of these findings extend beyond just theoretical astrophysics. By improving the efficiency and accuracy of galaxy evolution simulations, researchers can better understand the mechanisms governing cosmic structure formation and the lifecycle of stars. This knowledge has profound implications for fields such as cosmology, astrobiology, and even planetary sciences.

In summary, this pioneering study illustrates the remarkable potential of integrating artificial intelligence with traditional astrophysical modeling techniques. By overcoming computational bottlenecks, this research not only enhances our existing knowledge but also paves the way for future explorations into the cosmos. As technology continues to evolve, the insights gained from these simulations may ultimately help us answer some of the most pressing questions about our universe and our place within it.

Artificial intelligence has already begun its transformative journey through numerous industries, and now, its application in astrophysics demonstrates its utility in addressing complex scientific problems. As researchers continue to refine and improve these models, the possibilities for understanding astronomical phenomena appear virtually limitless.

The cosmic landscape offers a rich tapestry of mysteries, many of which remain unresolved. With this new AI-driven framework in play, we may soon witness a new era in astrophysics where the complexity of galaxy formation and evolution can be explored at unprecedented scales and resolutions. As we stand on the brink of this new frontier, the excitement within the scientific community is palpable, foreshadowing countless discoveries that lie ahead.

As the researchers embark on their journey to simulate a Milky Way-sized galaxy, the astronomical community will eagerly await the results, hopeful that this innovative study will illuminate uncharted territories of knowledge and potentially redefine our understanding of the universe.

Subject of Research: Machine learning in galaxy evolution simulations
Article Title: Pioneering AI-Driven Simulations of Galaxy Evolution: The Dawn of a New Era in Astrophysics
News Publication Date: [Publication Date Not Provided]
Web References: [Web References Not Provided]
References: [References Not Provided]
Image Credits: RIKEN

Keywords

Machine Learning, Galaxy Evolution, Supernova, Astrophysics, Numerical Simulations, High-Resolution Modeling, Artificial Intelligence, Cosmic Structure Formation, Milky Way, Star Formation