arXiv subjects

Francois Lanusse

Publications and source records attributed to Francois Lanusse.

At least 19 recordsLinked to original sources

arXiv2026

Emergent Transfer of a Physics Foundation Model from Simulation to Laboratory Turbulence

Whether physics foundation models can be usefully deployed on laboratory experiments remains an open question for scientific machine learning (ML). We test this question on the Rayleigh-Taylor instability (RTI), a ubiquitous and demanding fluid instability seen from tabletop flows to supernova explosions, in which small perturbations at a density interface grow into chaotic, multiscale mixing as a lighter fluid accelerates into a heavier one. Standard ML models struggle with RTI, and despite over a century of theoretical, numerical, and experimental work, it carries an unresolved discrepancy between simulation and experiment: the late-time mixing growth rate, $\alpha$, measured in most laboratory experiments ($\sim$ 0.06-0.07), is roughly three times the value from idealized direct numerical simulations (DNS, $\sim$ 0.02). The gap's origin remains debated. These properties make RTI a stringent test for a question that matters well beyond RTI: can foundation models trained only on simulations generalise to sparse, messy, and noisy laboratory settings? We finetune Walrus, a foundation model for continuum dynamics, on three or fewer DNS realizations and recover key RTI physics over long rollouts. Applied zero-shot to sliding-barrier laboratory data, the finetuned model leaves the DNS-like regime and enters the observed growth band, having never seen a single experimental sample. These results provide independent, data-driven evidence that initial conditions play a crucial role in the longstanding sim-experiment gap in $\alpha$. The model also generalises zero-shot to stable stratification, a buoyancy regime absent from training, correctly slowing mixing-layer growth. Together, our results show that foundation models can generalise well beyond their training data, predicting laboratory behavior and unseen physical regimes, opening new ways to probe longstanding simulation-experiment gaps.

physics.flu-dyn↗

arXiv2026

MIMIC: A Generative Multimodal Foundation Model for Biomolecules

Biological function emerges from coupled constraints across sequence, structure, regulation, evolution, and cellular context, yet most foundation models in biology are trained within one modality or for a fixed forward task. We present MIMIC, a generative multimodal foundation model trained on our newly curated and aligned dataset, LORE, linking nucleic acid, protein, evolutionary, structural, regulatory, and semantic/contextual modalities within partially observed biomolecular states. MIMIC uses a split-track encoder-decoder architecture to condition on arbitrary subsets of observed modalities and reconstruct or generate missing components of molecular state across the genome, transcriptome, and proteome. Multimodal conditioning consistently improves MIMIC's sequence reconstruction relative to sequence-only inputs, while its learned representations enable state-of-the-art performance on RNA and protein downstream tasks. MIMIC achieves state-of-the-art splicing prediction, and its joint generative formulation enables isoform-aware inference that further improves performance. Beyond prediction, the same generative framework supports constrained design. For RNA, MIMIC identifies corrective edits in a clinically relevant HBB splice-disrupting mutation without reverting it by using evolutionary and structural signals. For proteins, jointly conditioning on shape and surface chemistry of PD-L1 and hACE2 binding sites produces diverse, high-confidence sequences with strong in silico support for target binding. Finally, MIMIC uses experimental context as semantic conditioning to model assay-dependent RNA chemical probing, rather than treating context as a fixed output. Together, these results position MIMIC's aligned multimodal generative modeling as a strong foundation for unifying representation learning, conditional prediction, and constrained biomolecular design within a single model.

cs.AI↗

arXiv2025

Semantic search for 100M+ galaxy images using AI-generated captions

Finding scientifically interesting phenomena through slow manual labeling campaigns severely limits our ability to explore the billions of galaxy images produced by telescopes. In this work, we develop a pipeline to create a semantic search engine from completely unlabeled image data. Our method leverages Vision-Language Models (VLMs) to generate descriptions for galaxy images, then contrastively aligns a pre-trained astronomy foundation model with these embedded descriptions to produce searchable embeddings at scale. We find that current VLMs provide descriptions that are sufficiently informative to train a semantic search model that outperforms direct image similarity search. Our model, AION-Search, achieves state-of-the-art zero-shot performance on finding rare phenomena despite training on randomly selected images with no deliberate curation for rare cases. Furthermore, we introduce a VLM-based re-ranking method that nearly doubles the recall for our most challenging targets in the top-100 results. For the first time, AION-Search enables flexible semantic search for over 100 million galaxy images, enabling discovery from previously infeasible searches, including the identification of 36 new extragalactic stellar stream candidates. More broadly, our work provides an approach for making large, unlabeled scientific image archives semantically searchable, expanding data exploration capabilities in fields from Earth observation to microscopy. The code, data, and app are publicly available at https://github.com/NolanKoblischke/AION-Search

astro-ph.IM↗

arXiv2025

Walrus: A Cross-Domain Foundation Model for Continuum Dynamics

Foundation models have transformed machine learning for language and vision, but achieving comparable impact in physical simulation remains a challenge. Data heterogeneity and unstable long-term dynamics inhibit learning from sufficiently diverse dynamics, while varying resolutions and dimensionalities challenge efficient training on modern hardware. Through empirical and theoretical analysis, we incorporate new approaches to mitigate these obstacles, including a harmonic-analysis-based stabilization method, load-balanced distributed 2D and 3D training strategies, and compute-adaptive tokenization. Using these tools, we develop Walrus, a transformer-based foundation model developed primarily for fluid-like continuum dynamics. Walrus is pretrained on nineteen diverse scenarios spanning astrophysics, geoscience, rheology, plasma physics, acoustics, and classical fluids. Experiments show that Walrus outperforms prior foundation models on both short and long term prediction horizons on downstream tasks and across the breadth of pretraining data, while ablation studies confirm the value of our contributions to forecast stability, training throughput, and transfer performance over conventional approaches. Code and weights are released for community use.

cs.LG↗

arXiv2025

Bridging Simulators with Conditional Optimal Transport

We propose a new field-level emulator that bridges two simulators using unpaired simulation datasets. Our method leverages a flow-based approach to learn the likelihood transport from one simulator to the other. Since multiple transport maps exist, we employ Conditional Optimal Transport Flow Matching (COT-FM) to ensure that the transformation minimally distorts the underlying structure of the data. We demonstrate the effectiveness of this approach by bridging weak lensing simulators: a Lagrangian Perturbation Theory (LPT) to a N-body Particle-Mesh (PM). We demonstrate that our emulator captures the full correction between the simulators by showing that it enables full-field inference to accurately recover the true posterior, validating its accuracy beyond traditional summary statistics.

astro-ph.CO↗

arXiv2025

Universal Spectral Tokenization via Self-Supervised Panchromatic Representation Learning

Sequential scientific data span many resolutions and domains, and unifying them into a common representation is a key step toward developing foundation models for the sciences. Astronomical spectra exemplify this challenge: massive surveys have collected millions of spectra across a wide range of wavelengths and resolutions, yet analyses remain fragmented across spectral domains (e.g., optical vs. infrared) and object types (e.g., stars vs. galaxies), limiting the ability to pool information across datasets. We present a deep learning model that jointly learns from heterogeneous spectra in a self-supervised manner. Our universal spectral tokenizer processes spectra from a variety of object types and resolutions directly on their native wavelength grids, producing intrinsically aligned, homogeneous, and physically meaningful representations that can be efficiently adapted to achieve competitive performance across a range of downstream tasks. For the first time, we demonstrate that a single model can unify spectral data across resolutions and domains, suggesting that our model can serve as a powerful building block for foundation models in astronomy -- and potentially extend to other scientific domains with heterogeneous sequential data, such as climate and healthcare.

astro-ph.IM↗

arXiv2025

AION-1: Omnimodal Foundation Model for Astronomical Sciences

While foundation models have shown promise across a variety of fields, astronomy still lacks a unified framework for joint modeling across its highly diverse data modalities. In this paper, we present AION-1, a family of large-scale multimodal foundation models for astronomy. AION-1 integrates heterogeneous imaging, spectroscopic, and scalar data using a two-stage architecture: modality-specific tokenization followed by transformer-based masked modeling of cross-modal token sequences. The model is pretrained on five large-scale surveys: Legacy Survey, Hyper Suprime-Cam (HSC), Sloan Digital Sky Survey (SDSS), Dark Energy Spectroscopic Instrument (DESI), and Gaia. These span more than 200 million observations of stars, galaxies, and quasars. With a single frozen encoder, AION-1 achieves strong results on a broad suite of downstream tasks, including galaxy and stellar property estimation, galaxy morphology classification, similarity-based retrieval, galaxy image segmentation, and spectral super-resolution. We release AION-1 model variants ranging from 300 M to 3.1 B parameters. Beyond astronomy, AION-1 provides a scalable blueprint for multimodal scientific foundation models that can seamlessly integrate noisy, instrument-specific observations. All code, tokenizers, pretrained weights, and a lightweight evaluation suite are released under an open-source license.

astro-ph.IM↗

arXiv2025

Measuring the intracluster light fraction with machine learning

The intracluster light (ICL) is an important tracer of a galaxy cluster's history and past interactions. However, only small samples have been studied to date due to its very low surface brightness and the heavy manual involvement required for the majority of measurement algorithms. Upcoming large imaging surveys such as the Vera C. Rubin Observatory's Legacy Survey of Space and Time are expected to vastly expand available samples of deep cluster images. However, to process this increased amount of data, we need faster, fully automated methods to streamline the measurement process. This paper presents a machine learning model designed to automatically measure the ICL fraction in large samples of images, with no manual preprocessing required. We train the fully supervised model on a training dataset of 50,000 images with injected artificial ICL profiles. We then transfer its learning onto real data by fine-tuning with a sample of 101 real clusters with their ICL fraction measured manually using the surface brightness threshold method. With this process, the model is able to effectively learn the task and then adapt its learning to real cluster images. Our model can be directly applied to Hyper Suprime-Cam images, processing up to 500 images in a matter of seconds on a single GPU, or fine-tuned for other imaging surveys such as LSST, with the fine-tuning process taking just 3 minutes. The model could also be retrained to match other ICL measurement methods. Our model and the code for training it is made available on GitHub.

astro-ph.GA↗

arXiv2024

The Multimodal Universe: Enabling Large-Scale Machine Learning with 100TB of Astronomical Scientific Data

We present the MULTIMODAL UNIVERSE, a large-scale multimodal dataset of scientific astronomical data, compiled specifically to facilitate machine learning research. Overall, the MULTIMODAL UNIVERSE contains hundreds of millions of astronomical observations, constituting 100\,TB of multi-channel and hyper-spectral images, spectra, multivariate time series, as well as a wide variety of associated scientific measurements and "metadata". In addition, we include a range of benchmark tasks representative of standard practices for machine learning methods in astrophysics. This massive dataset will enable the development of large multi-modal models specifically targeted towards scientific applications. All codes used to compile the MULTIMODAL UNIVERSE and a description of how to access the data is available at https://github.com/MultimodalUniverse/MultimodalUniverse

astro-ph.IM↗

arXiv2024

Geometric deep learning for galaxy-halo connection: a case study for galaxy intrinsic alignments

Forthcoming cosmological imaging surveys, such as the Rubin Observatory LSST, require large-scale simulations encompassing realistic galaxy populations for a variety of scientific applications. Of particular concern is the phenomenon of intrinsic alignments (IA), whereby galaxies orient themselves towards overdensities, potentially introducing significant systematic biases in weak gravitational lensing analyses if they are not properly modeled. Due to computational constraints, simulating the intricate details of galaxy formation and evolution relevant to IA across vast volumes is impractical. As an alternative, we propose a Deep Generative Model trained on the IllustrisTNG-100 simulation to sample 3D galaxy shapes and orientations to accurately reproduce intrinsic alignments along with correlated scalar features. We model the cosmic web as a set of graphs, each graph representing a halo with nodes representing the subhalos/galaxies. The architecture consists of a SO(3) $\times$ $\mathbb{R}^n$ diffusion generative model, for galaxy orientations and $n$ scalars, implemented with E(3) equivariant Graph Neural Networks that explicitly respect the Euclidean symmetries of our Universe. The model is able to learn and predict features such as galaxy orientations that are statistically consistent with the reference simulation. Notably, our model demonstrates the ability to jointly model Euclidean-valued scalars (galaxy sizes, shapes, and colors) along with non-Euclidean valued SO(3) quantities (galaxy orientations) that are governed by highly complex galactic physics at non-linear scales.

astro-ph.GA↗

arXiv2024

Teaching dark matter simulations to speak the halo language

We develop a transformer-based conditional generative model for discrete point objects and their properties. We use it to build a model for populating cosmological simulations with gravitationally collapsed structures called dark matter halos. Specifically, we condition our model with dark matter distribution obtained from fast, approximate simulations to recover the correct three-dimensional positions and masses of individual halos. This leads to a first model that can recover the statistical properties of the halos at small scales to better than 3% level using an accelerated dark matter simulation. This trained model can then be applied to simulations with significantly larger volumes which would otherwise be computationally prohibitive with traditional simulations, and also provides a crucial missing link in making end-to-end differentiable cosmological simulations. The code, named GOTHAM (Generative cOnditional Transformer for Halo's Auto-regressive Modeling) is publicly available at \url{https://github.com/shivampcosmo/GOTHAM}.

astro-ph.CO↗

arXiv2023

Unified framework for diffusion generative models in SO(3): applications in computer vision and astrophysics

Diffusion-based generative models represent the current state-of-the-art for image generation. However, standard diffusion models are based on Euclidean geometry and do not translate directly to manifold-valued data. In this work, we develop extensions of both score-based generative models (SGMs) and Denoising Diffusion Probabilistic Models (DDPMs) to the Lie group of 3D rotations, SO(3). SO(3) is of particular interest in many disciplines such as robotics, biochemistry and astronomy/cosmology science. Contrary to more general Riemannian manifolds, SO(3) admits a tractable solution to heat diffusion, and allows us to implement efficient training of diffusion models. We apply both SO(3) DDPMs and SGMs to synthetic densities on SO(3) and demonstrate state-of-the-art results. Additionally, we demonstrate the practicality of our model on pose estimation tasks and in predicting correlated galaxy orientations for astrophysics/cosmology.

cs.LG↗

arXiv2023

An Empirical Model For Intrinsic Alignments: Insights From Cosmological Simulations

We extend current models of the halo occupation distribution (HOD) to include a flexible, empirical framework for the forward modeling of the intrinsic alignment (IA) of galaxies. A primary goal of this work is to produce mock galaxy catalogs for the purpose of validating existing models and methods for the mitigation of IA in weak lensing measurements. This technique can also be used to produce new, simulation-based predictions for IA and galaxy clustering. Our model is probabilistically formulated, and rests upon the assumption that the orientations of galaxies exhibit a correlation with their host dark matter (sub)halo orientation or with their position within the halo. We examine the necessary components and phenomenology of such a model by considering the alignments between (sub)halos in a cosmological dark matter only simulation. We then validate this model for a realistic galaxy population in a set of simulations in the IllustrisTNG suite. We create an HOD mock with TNG-like correlations using our method, constraining the associated IA model parameters, with the $\chi^2_{\rm dof}$ between our model's correlations and those of Illustris matching as closely as 1.4 and 1.1 for orientation--position and orientation--orientation correlation functions, respectively. By modeling the misalignment between galaxies and their host halo, we show that the 3-dimensional two-point position and orientation correlation functions of simulated (sub)halos and galaxies can be accurately reproduced from quasi-linear scales down to $0.1~h^{-1}{\rm Mpc}$. We also find evidence for environmental influence on IA within a halo. Our publicly-available software provides a key component enabling efficient determination of Bayesian posteriors on IA model parameters using observational measurements of galaxy-orientation correlation functions in the highly nonlinear regime.

astro-ph.CO↗

arXiv2023

xVal: A Continuous Numerical Tokenization for Scientific Language Models

Due in part to their discontinuous and discrete default encodings for numbers, Large Language Models (LLMs) have not yet been commonly used to process numerically-dense scientific datasets. Rendering datasets as text, however, could help aggregate diverse and multi-modal scientific data into a single training corpus, thereby potentially facilitating the development of foundation models for science. In this work, we introduce xVal, a strategy for continuously tokenizing numbers within language models that results in a more appropriate inductive bias for scientific applications. By training specially-modified language models from scratch on a variety of scientific datasets formatted as text, we find that xVal generally outperforms other common numerical tokenization strategies on metrics including out-of-distribution generalization and computational efficiency.

stat.ML↗

arXiv2023

Multiple Physics Pretraining for Physical Surrogate Models

We introduce multiple physics pretraining (MPP), an autoregressive task-agnostic pretraining approach for physical surrogate modeling of spatiotemporal systems with transformers. In MPP, rather than training one model on a specific physical system, we train a backbone model to predict the dynamics of multiple heterogeneous physical systems simultaneously in order to learn features that are broadly useful across systems and facilitate transfer. In order to learn effectively in this setting, we introduce a shared embedding and normalization strategy that projects the fields of multiple systems into a shared embedding space. We validate the efficacy of our approach on both pretraining and downstream tasks over a broad fluid mechanics-oriented benchmark. We show that a single MPP-pretrained transformer is able to match or outperform task-specific baselines on all pretraining sub-tasks without the need for finetuning. For downstream tasks, we demonstrate that finetuning MPP-trained models results in more accurate predictions across multiple time-steps on systems with previously unseen physical components or higher dimensional systems compared to training from scratch or finetuning pretrained video foundation models. We open-source our code and model weights trained at multiple scales for reproducibility.

cs.LG↗

arXiv2023

AstroCLIP: A Cross-Modal Foundation Model for Galaxies

We present AstroCLIP, a single, versatile model that can embed both galaxy images and spectra into a shared, physically meaningful latent space. These embeddings can then be used - without any model fine-tuning - for a variety of downstream tasks including (1) accurate in-modality and cross-modality semantic similarity search, (2) photometric redshift estimation, (3) galaxy property estimation from both images and spectra, and (4) morphology classification. Our approach to implementing AstroCLIP consists of two parts. First, we embed galaxy images and spectra separately by pretraining separate transformer-based image and spectrum encoders in self-supervised settings. We then align the encoders using a contrastive loss. We apply our method to spectra from the Dark Energy Spectroscopic Instrument and images from its corresponding Legacy Imaging Survey. Overall, we find remarkable performance on all downstream tasks, even relative to supervised baselines. For example, for a task like photometric redshift prediction, we find similar performance to a specifically-trained ResNet18, and for additional tasks like physical property estimation (stellar mass, age, metallicity, and sSFR), we beat this supervised baseline by 19\% in terms of $R^2$. We also compare our results to a state-of-the-art self-supervised single-modal model for galaxy images, and find that our approach outperforms this benchmark by roughly a factor of two on photometric redshift estimation and physical property prediction in terms of $R^2$, while remaining roughly in-line in terms of morphology classification. Ultimately, our approach represents the first cross-modal self-supervised model for galaxies, and the first self-supervised transformer-based architectures for galaxy images and spectra.

astro-ph.IM↗

arXiv2023

Detecting Galaxy Tidal Features Using Self-Supervised Representation Learning

Low surface brightness substructures around galaxies, known as tidal features, are a valuable tool in the detection of past or ongoing galaxy mergers, and their properties can answer questions about the progenitor galaxies involved in the interactions. The assembly of current tidal feature samples is primarily achieved using visual classification, making it difficult to construct large samples and draw accurate and statistically robust conclusions about the galaxy evolution process. With upcoming large optical imaging surveys such as the Vera C. Rubin Observatory Legacy Survey of Space and Time (LSST), predicted to observe billions of galaxies, it is imperative that we refine our methods of detecting and classifying samples of merging galaxies. This paper presents promising results from a self-supervised machine learning model, trained on data from the Ultradeep layer of the Hyper Suprime-Cam Subaru Strategic Program optical imaging survey, designed to automate the detection of tidal features. We find that self-supervised models are capable of detecting tidal features, and that our model outperforms previous automated tidal feature detection methods, including a fully supervised model. An earlier method applied to real galaxy images achieved 76% completeness for 22% contamination, while our model achieves considerably higher (96%) completeness for the same level of contamination. We emphasise a number of advantages of self-supervised models over fully supervised models including maintaining excellent performance when using only 50 labelled examples for training, and the ability to perform similarity searches using a single example of a galaxy with tidal features.

astro-ph.GA↗

arXiv2023

Detecting Tidal Features using Self-Supervised Representation Learning

Low surface brightness substructures around galaxies, known as tidal features, are a valuable tool in the detection of past or ongoing galaxy mergers. Their properties can answer questions about the progenitor galaxies involved in the interactions. This paper presents promising results from a self-supervised machine learning model, trained on data from the Ultradeep layer of the Hyper Suprime-Cam Subaru Strategic Program optical imaging survey, designed to automate the detection of tidal features. We find that self-supervised models are capable of detecting tidal features and that our model outperforms previous automated tidal feature detection methods, including a fully supervised model. The previous state of the art method achieved 76% completeness for 22% contamination, while our model achieves considerably higher (96%) completeness for the same level of contamination.

astro-ph.GA↗