Searcharxiv⌕ Search

arXiv subjects

Pablo Lemos

Publications and source records attributed to Pablo Lemos.

At least 19 recordsLinked to original sources
arXiv2026

MIRA: A Score for Conditional Distribution Accuracy and Model Comparison

We introduce Mira, a sample-based score for assessing the accuracy of a candidate conditional distribution using only joint samples from the true data-generating process. Relying on the principle that distributions coincide if they assign equal probability mass to all regions, we derive an analytic expression for the Mira statistic, whose average defines the Mira score. This formulation further allows us to compute theoretical reference values and uncertainty estimates when the candidate distribution matches the true one. This framework enables model comparison by quantifying the alignment between the conditional distribution of a candidate model and the true data generating process. Consequently, Mira enables Bayesian model comparison through direct posterior validation, bypassing the challenging evidence computation. We demonstrate its effectiveness across several toy problems and Bayesian inference tasks.

stat.ML↗
arXiv2026

Introducing sapphire: Towards Hybrid Physics-Informed, Data-Driven Modeling of Galaxy Formation

Semi-analytic models (SAMs) have been treating galaxy populations as dynamical systems for $\gtrsim50$ years, but their evolution equations remain poorly constrained. We introduce sapphire, a modular, automatically differentiable, GPU-accelerated SAM written in JAX. For the first time, we compute exact Jacobian and Hessian matrices of a galaxy formation SAM, using the Pandya et al. (2023) nonlinear differential equation system as an example. These allow efficient, interpretable local and global sensitivity analyses, which reveal that supernova energy loading is the key astrophysical parameter. We use gradient descent and Hamiltonian Monte Carlo (HMC) to perform comprehensive mock parameter recovery tests. These indicate that the $z=0$ stellar-to-halo-mass relation alone does not contain enough information to infer many astrophysical parameters. Using observations of star-forming galaxies from the MaNGA survey and the Behroozi et al. (2019) empirical model as one baseline, we derive multiple posteriors assuming different combinations of data, including $z=0$ interstellar medium gas fractions and metallicities. The inferred physical parameters suggest that galaxies self-regulate their star formation primarily through preventative rather than ejective feedback, though this remains uncertain due to the lack of satellite galaxies, black holes and multi-phase galactic atmosphere physics. Both Fisher and HMC forecasts demonstrate the potential of sapphire to enable precision inference for galaxy formation and cosmology in a hybrid physics-informed, data-driven way, but more work is needed to expand its library of models and methods. We make sapphire publicly available at https://github.com/virajpandya/sapphire.

astro-ph.GA↗
arXiv2026

A Fast Generative Framework for High-dimensional Posterior Sampling: Application to CMB Delensing

We introduce a deep generative framework for high-dimensional Bayesian inference that enables efficient posterior sampling. As telescopes and simulations rapidly expand the volume and resolution of astrophysical data, fast simulation-based inference methods are increasingly needed to extract scientific insights. While diffusion-based approaches offer high-quality generative capabilities, they are hindered by slow sampling speeds. Our method performs posterior sampling an order of magnitude faster than a diffusion baseline. Applied to the problem of CMB delensing, it successfully recovers the unlensed CMB power spectrum from simulated observations. The model also remains robust to shifts in cosmological parameters, demonstrating its potential for out-of-distribution generalization and application to observational cosmological data.

astro-ph.IM↗
arXiv2025

Solving Bayesian inverse problems with diffusion priors and off-policy RL

This paper presents a practical application of Relative Trajectory Balance (RTB), a recently introduced off-policy reinforcement learning (RL) objective that can asymptotically solve Bayesian inverse problems optimally. We extend the original work by using RTB to train conditional diffusion model posteriors from pretrained unconditional priors for challenging linear and non-linear inverse problems in vision, and science. We use the objective alongside techniques such as off-policy backtracking exploration to improve training. Importantly, our results show that existing training-free diffusion posterior methods struggle to perform effective posterior inference in latent space due to inherent biases.

cs.LG↗
arXiv2024

Causal Discovery in Astrophysics: Unraveling Supermassive Black Hole and Galaxy Coevolution

Correlation does not imply causation, but patterns of statistical association between variables can be exploited to infer a causal structure (even with purely observational data) with the burgeoning field of causal discovery. As a purely observational science, astrophysics has much to gain by exploiting these new methods. The supermassive black hole (SMBH)--galaxy interaction has long been constrained by observed scaling relations, that is low-scatter correlations between variables such as SMBH mass and the central velocity dispersion of stars in a host galaxy's bulge. This study, using advanced causal discovery techniques and an up-to-date dataset, reveals a causal link between galaxy properties and dynamically-measured SMBH masses. We apply a score-based Bayesian framework to compute the exact conditional probabilities of every causal structure that could possibly describe our galaxy sample. With the exact posterior distribution, we determine the most likely causal structures and notice a probable causal reversal when separating galaxies by morphology. In elliptical galaxies, bulge properties (built from major mergers) tend to influence SMBH growth, while in spiral galaxies, SMBHs are seen to affect host galaxy properties, potentially through feedback in gas-rich environments. For spiral galaxies, SMBHs progressively quench star formation, whereas in elliptical galaxies, quenching is complete, and the causal connection has reversed. Our findings support theoretical models of hierarchical assembly of galaxies and active galactic nuclei feedback regulating galaxy evolution. Our study suggests the potentiality for further exploration of causal links in astrophysical and cosmological scaling relations, as well as any other observational science.

astro-ph.GA↗
arXiv2024

Amortizing intractable inference in diffusion models for vision, language, and control

Diffusion models have emerged as effective distribution estimators in vision, language, and reinforcement learning, but their use as priors in downstream tasks poses an intractable posterior inference problem. This paper studies amortized sampling of the posterior over data, $\mathbf{x}\sim p^{\rm post}(\mathbf{x})\propto p(\mathbf{x})r(\mathbf{x})$, in a model that consists of a diffusion generative model prior $p(\mathbf{x})$ and a black-box constraint or likelihood function $r(\mathbf{x})$. We state and prove the asymptotic correctness of a data-free learning objective, relative trajectory balance, for training a diffusion model that samples from this posterior, a problem that existing methods solve only approximately or in restricted cases. Relative trajectory balance arises from the generative flow network perspective on diffusion models, which allows the use of deep reinforcement learning techniques to improve mode coverage. Experiments illustrate the broad potential of unbiased inference of arbitrary posteriors under diffusion priors: in vision (classifier guidance), language (infilling under a discrete diffusion LLM), and multimodal data (text-to-image generation). Beyond generative modeling, we apply relative trajectory balance to the problem of continuous control with a score-based behavior prior, achieving state-of-the-art results on benchmarks in offline reinforcement learning.

cs.LG↗
arXiv2024

Sequence-Augmented SE(3)-Flow Matching For Conditional Protein Backbone Generation

Proteins are essential for almost all biological processes and derive their diverse functions from complex 3D structures, which are in turn determined by their amino acid sequences. In this paper, we exploit the rich biological inductive bias of amino acid sequences and introduce FoldFlow-2, a novel sequence-conditioned SE(3)-equivariant flow matching model for protein structure generation. FoldFlow-2 presents substantial new architectural features over the previous FoldFlow family of models including a protein large language model to encode sequence, a new multi-modal fusion trunk that combines structure and sequence representations, and a geometric transformer based decoder. To increase diversity and novelty of generated samples -- crucial for de-novo drug design -- we train FoldFlow-2 at scale on a new dataset that is an order of magnitude larger than PDB datasets of prior works, containing both known proteins in PDB and high-quality synthetic structures achieved through filtering. We further demonstrate the ability to align FoldFlow-2 to arbitrary rewards, e.g. increasing secondary structures diversity, by introducing a Reinforced Finetuning (ReFT) objective. We empirically observe that FoldFlow-2 outperforms previous state-of-the-art protein structure-based generative models, improving over RFDiffusion in terms of unconditional generation across all metrics including designability, diversity, and novelty across all protein lengths, as well as exhibiting generalization on the task of equilibrium conformation sampling. Finally, we demonstrate that a fine-tuned FoldFlow-2 makes progress on challenging conditional design tasks such as designing scaffolds for the VHH nanobody.

cs.LG↗
arXiv2024

Ejection and Dynamics of Aggregates in the Coma of Comet 67P/Churyumov-Gerasimenko

The process of cometary activity continues to pose a challenging question in cometary science. The activity modeling of comet 67P/Churyumov-Gerasimenko, based on data from the Rosetta mission, has significantly enhanced our comprehension of cometary activity. But thermophysical models have difficulties in simultaneously explaining the production rates of various gas species and dust. It has been suggested that different gas species might be responsible for the ejection of refractory material in distinct size ranges. This work focuses on investigating abundance and the ejection mechanisms of large ($\gtrsim$ 1 cm) aggregates from the comet nucleus. We aim to determine their properties and map the distribution of their source regions across the comet surface. This can place constraints on activity models for comets. We examined 189 images acquired at five epochs by the OSIRIS/NAC instrument. Our goal was to identify bright tracks produced by individual aggregates as they traversed the camera field of view. We generated synthetic images based on the output of dynamical simulations involving various types of aggregates. By comparing these synthetic images with the observations, we determine the characteristics of the simulated aggregates that most closely resembled the observations. We identified over 30000 tracks present in the OSIRIS images, derived constraints on the characteristics of the aggregates and mapped their origins on the nucleus surface. The aggregates have an average radius of $\simeq5$ cm, and a bulk density consistent with that of the comet's nucleus. Due to their size, gas drag exerts only a minor influence on their dynamical behavior, so an initial velocity is needed in order to bring them into the camera field of view. The source regions of these aggregates are predominantly located near the boundaries of distinct terrains on the surface.

astro-ph.EP↗
arXiv2024

{\sc SimBIG}: Cosmological Constraints using Simulation-Based Inference of Galaxy Clustering with Marked Power Spectra

We present the first $\Lambda$CDM cosmological analysis performed on a galaxy survey using marked power spectra. The marked power spectrum is the two-point function of a marked field, where galaxies are weighted by a function that depends on their local density. The presence of the mark leads these statistics to contain higher-order information of the original galaxy field, making them a good candidate to exploit the non-Gaussian information of a galaxy catalog. In this work we make use of \simbig, a forward modeling framework for galaxy clustering analyses, and perform simulation-based inference using normalizing flows to infer the posterior distribution of the $\Lambda$CDM cosmological parameters. We consider different mark configurations (ways to weight the galaxy field) and deploy them in the \simbig~pipeline to analyze the corresponding marked power spectra measured from a subset of the BOSS galaxy sample. We analyze the redshift-space mark power spectra decomposed in $\ell = 0, 2, 4$ multipoles and include scales up to the non-linear regime. Among the various mark configurations considered, the ones that give the most stringent cosmological constraints produce posterior median and $68\%$ confidence limits on the growth of structure parameters equal to $\Omega_m=0.273^{+0.040}_{-0.030}$ and $\sigma_8=0.777^{+0.077}_{-0.071}$. Compared to a perturbation theory analysis using the power spectrum of the same dataset, the \simbig~marked power spectra constraints on $\sigma_8$ are up to $1.2\times$ tighter, while no improvement is seen for the other cosmological parameters.

astro-ph.CO↗
arXiv2024

Dynamics and potential origins of decimeter-sized particles around comet 67P/Churyumov-Gerasimenko

Methods. We algorithmically tracked thousands of individual particles through four OSIRIS/NAC image sequences of 67P's near-nucleus coma. We then traced concentrated particle groups back to the nucleus surface, and estimated their potential source regions, size distributions, and projected dynamical parameters. Finally, we compared the observed activity to dust coma simulations. Results. We traced back 409 decimeter-sized particles to four suspected source regions. The regions strongly overlap and are mostly confined to the Khonsu-Atum-Anubis area. The activity may be linked to rugged terrain, and the erosion of fine dust and the ejection of large boulders may be mutually exclusive. Power-law indices fitted to the particle size--frequency distributions range from $3.4 \pm 0.3$ to $3.8 \pm 0.4$. Gas drag fits to the radial particle accelerations provide an estimate for the local gas production rates ($Q_\text{g} = 3.6 \cdot 10^{-5}$ kg s$^{-1}$ m$^{-2}$), which is several times higher than our model predictions based on purely insolation-driven water ice sublimation. Our observational results and our modeling results both reveal that our particles were likely ejected with substantial nonzero initial velocities of around 0.5$-$0.6 m s$^{-1}$. Conclusions. Our findings strongly suggest that the observed ejection of decimeter-sized particles cannot be explained by water ice sublimation and favorable illumination conditions alone. Instead, the local structures and compositions of the source regions likely play a major role. In line with current ejection models of decimeter-sized particles, we deem an overabundance of CO$_2$ ice and its sublimation to be the most probable driver. In addition, because of the significant initial velocities, we suspect the ejection events to be considerably more energetic than gradual liftoffs.

astro-ph.EP↗
arXiv2024

Iterated Denoising Energy Matching for Sampling from Boltzmann Densities

Efficiently generating statistically independent samples from an unnormalized probability distribution, such as equilibrium samples of many-body systems, is a foundational problem in science. In this paper, we propose Iterated Denoising Energy Matching (iDEM), an iterative algorithm that uses a novel stochastic score matching objective leveraging solely the energy function and its gradient -- and no data samples -- to train a diffusion-based sampler. Specifically, iDEM alternates between (I) sampling regions of high model density from a diffusion-based sampler and (II) using these samples in our stochastic matching objective to further improve the sampler. iDEM is scalable to high dimensions as the inner matching objective, is simulation-free, and requires no MCMC samples. Moreover, by leveraging the fast mode mixing behavior of diffusion, iDEM smooths out the energy landscape enabling efficient exploration and learning of an amortized sampler. We evaluate iDEM on a suite of tasks ranging from standard synthetic energy functions to invariant $n$-body particle systems. We show that the proposed approach achieves state-of-the-art performance on all metrics and trains $2-5\times$ faster, which allows it to be the first method to train using energy on the challenging $55$-particle Lennard-Jones system.

cs.LG↗
arXiv2024

Improved off-policy training of diffusion samplers

We study the problem of training diffusion models to sample from a distribution with a given unnormalized density or energy function. We benchmark several diffusion-structured inference methods, including simulation-based variational approaches and off-policy methods (continuous generative flow networks). Our results shed light on the relative advantages of existing algorithms while bringing into question some claims from past work. We also propose a novel exploration strategy for off-policy methods, based on local search in the target space with the use of a replay buffer, and show that it improves the quality of samples on a variety of target distributions. Our code for the sampling methods and benchmarks studied is made public at https://github.com/GFNOrg/gfn-diffusion as a base for future work on diffusion models for amortized inference.

cs.LG↗
arXiv2024

PQMass: Probabilistic Assessment of the Quality of Generative Models using Probability Mass Estimation

We propose a likelihood-free method for comparing two distributions given samples from each, with the goal of assessing the quality of generative models. The proposed approach, PQMass, provides a statistically rigorous method for assessing the performance of a single generative model or the comparison of multiple competing models. PQMass divides the sample space into non-overlapping regions and applies chi-squared tests to the number of data samples that fall within each region, giving a p-value that measures the probability that the bin counts derived from two sets of samples are drawn from the same multinomial distribution. PQMass does not depend on assumptions regarding the density of the true distribution, nor does it rely on training or fitting any auxiliary models. We evaluate PQMass on data of various modalities and dimensions, demonstrating its effectiveness in assessing the quality, novelty, and diversity of generated samples. We further show that PQMass scales well to moderately high-dimensional data and thus obviates the need for feature extraction in practical applications.

stat.ML↗
arXiv2024

LtU-ILI: An All-in-One Framework for Implicit Inference in Astrophysics and Cosmology

This paper presents the Learning the Universe Implicit Likelihood Inference (LtU-ILI) pipeline, a codebase for rapid, user-friendly, and cutting-edge machine learning (ML) inference in astrophysics and cosmology. The pipeline includes software for implementing various neural architectures, training schemata, priors, and density estimators in a manner easily adaptable to any research workflow. It includes comprehensive validation metrics to assess posterior estimate coverage, enhancing the reliability of inferred results. Additionally, the pipeline is easily parallelizable and is designed for efficient exploration of modeling hyperparameters. To demonstrate its capabilities, we present real applications across a range of astrophysics and cosmology problems, such as: estimating galaxy cluster masses from X-ray photometry; inferring cosmology from matter power spectra and halo point clouds; characterizing progenitors in gravitational wave signals; capturing physical dust parameters from galaxy colors and luminosities; and establishing properties of semi-analytic models of galaxy formation. We also include exhaustive benchmarking and comparisons of all implemented methods as well as discussions about the challenges and pitfalls of ML inference in astronomical sciences. All code and examples are made publicly available at https://github.com/maho3/ltu-ili.

astro-ph.IM↗
arXiv2024

${\rm S{\scriptsize IM}BIG}$: Cosmological Constraints from the Redshift-Space Galaxy Skew Spectra

Extracting the non-Gaussian information of the cosmic large-scale structure (LSS) is vital in unlocking the full potential of the rich datasets from the upcoming stage-IV galaxy surveys. Galaxy skew spectra serve as efficient beyond-two-point statistics, encapsulating essential bispectrum information with computational efficiency akin to power spectrum analysis. This paper presents the first cosmological constraints from analyzing the full set of redshift-space galaxy skew spectra of the data from the SDSS-III BOSS, accessing cosmological information down to nonlinear scales. Employing the ${\rm S{\scriptsize IM}BIG}$ forward modeling framework and simulation-based inference via normalizing flows, we analyze the CMASS-SGC sub-sample, which constitute approximately 10\% of the full BOSS data. Analyzing the scales up to $k_{\rm max}=0.5 \, {\rm Mpc}^{-1}h$, we find that the skew spectra improve the constraints on $\Omega_{\rm m}, \Omega_{\rm b}, h$, and $n_s$ by 34\%, 35\%, 18\%, 10\%, respectively, compared to constraints from previous ${\rm S{\scriptsize IM}BIG}$ power spectrum multipoles analysis, yielding $\Omega_{\rm m}=0.288^{+0.024}_{-0.034}$, $\Omega_{\rm b}= 0.043^{+0.005}_{-0.007}$, $h=0.759^{+0.104}_{-0.050}$, $n_{\rm s} = 0.918^{+0.041}_{-0.090}$ (at 68\% confidence limit). On the other hand, the constraints on $\sigma_8$ are weaker than from the power spectrum. Including the Big Bang Nucleosynthesis (BBN) prior on baryon density reduces the uncertainty on the Hubble parameter further, achieving $h=0.750^{+0.034}_{-0.032}$, which is a 38\% improvement over the constraint from the power spectrum with the same prior. Compared to the ${\rm S{\scriptsize IM}BIG}$ bispectrum (monopole) analysis, skew spectra offer comparable constraints on larger scales ($k_{\rm max}<0.3\, {\rm Mpc}^{-1}h$) for most parameters except for $\sigma_8$.

astro-ph.CO↗
arXiv2023

Improving Gradient-guided Nested Sampling for Posterior Inference

We present a performant, general-purpose gradient-guided nested sampling algorithm, ${\tt GGNS}$, combining the state of the art in differentiable programming, Hamiltonian slice sampling, clustering, mode separation, dynamic nested sampling, and parallelization. This unique combination allows ${\tt GGNS}$ to scale well with dimensionality and perform competitively on a variety of synthetic and real-world problems. We also show the potential of combining nested sampling with generative flow networks to obtain large amounts of high-quality samples from the posterior distribution. This combination leads to faster mode discovery and more accurate estimates of the partition function.

cs.LG↗
arXiv2023

Bayesian Imaging for Radio Interferometry with Score-Based Priors

The inverse imaging task in radio interferometry is a key limiting factor to retrieving Bayesian uncertainties in radio astronomy in a computationally effective manner. We use a score-based prior derived from optical images of galaxies to recover images of protoplanetary disks from the DSHARP survey. We demonstrate that our method produces plausible posterior samples despite the misspecified galaxy prior. We show that our approach produces results which are competitive with existing radio interferometry imaging algorithms.

astro-ph.IM↗
arXiv2023

Learning an Effective Evolution Equation for Particle-Mesh Simulations Across Cosmologies

Particle-mesh simulations trade small-scale accuracy for speed compared to traditional, computationally expensive N-body codes in cosmological simulations. In this work, we show how a data-driven model could be used to learn an effective evolution equation for the particles, by correcting the errors of the particle-mesh potential incurred on small scales during simulations. We find that our learnt correction yields evolution equations that generalize well to new, unseen initial conditions and cosmologies. We further demonstrate that the resulting corrected maps can be used in a simulation-based inference framework to yield an unbiased inference of cosmological parameters. The model, a network implemented in Fourier space, is exclusively trained on the particle positions and velocities.

astro-ph.CO↗