FlowLensing: Simulating Gravitational Lensing with Flow Matching

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Sayed, Hamees, Reddy, Pranath, Toomey, Michael W., Gleyzer, Sergei
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912708517953536
author Sayed, Hamees
Reddy, Pranath
Toomey, Michael W.
Gleyzer, Sergei
author_facet Sayed, Hamees
Reddy, Pranath
Toomey, Michael W.
Gleyzer, Sergei
contents Gravitational lensing is one of the most powerful probes of dark matter, yet creating high-fidelity lensed images at scale remains a bottleneck. Existing tools rely on ray-tracing or forward-modeling pipelines that, while precise, are prohibitively slow. We introduce FlowLensing, a Diffusion Transformer-based compact and efficient flow-matching model for strong gravitational lensing simulation. FlowLensing operates in both discrete and continuous regimes, handling classes such as different dark matter models as well as continuous model parameters ensuring physical consistency. By enabling scalable simulations, our model can advance dark matter studies, specifically for probing dark matter substructure in cosmological surveys. We find that our model achieves a speedup of over 200$\times$ compared to classical simulators for intensive dark matter models, with high fidelity and low inference latency. FlowLensing enables rapid, scalable, and physically consistent image synthesis, offering a practical alternative to traditional forward-modeling pipelines.
format Preprint
id arxiv_https___arxiv_org_abs_2510_07878
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle FlowLensing: Simulating Gravitational Lensing with Flow Matching
Sayed, Hamees
Reddy, Pranath
Toomey, Michael W.
Gleyzer, Sergei
Instrumentation and Methods for Astrophysics
Computer Vision and Pattern Recognition
Gravitational lensing is one of the most powerful probes of dark matter, yet creating high-fidelity lensed images at scale remains a bottleneck. Existing tools rely on ray-tracing or forward-modeling pipelines that, while precise, are prohibitively slow. We introduce FlowLensing, a Diffusion Transformer-based compact and efficient flow-matching model for strong gravitational lensing simulation. FlowLensing operates in both discrete and continuous regimes, handling classes such as different dark matter models as well as continuous model parameters ensuring physical consistency. By enabling scalable simulations, our model can advance dark matter studies, specifically for probing dark matter substructure in cosmological surveys. We find that our model achieves a speedup of over 200$\times$ compared to classical simulators for intensive dark matter models, with high fidelity and low inference latency. FlowLensing enables rapid, scalable, and physically consistent image synthesis, offering a practical alternative to traditional forward-modeling pipelines.
title FlowLensing: Simulating Gravitational Lensing with Flow Matching
topic Instrumentation and Methods for Astrophysics
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2510.07878