Papers
Topics
Authors
Recent
Assistant
AI Research Assistant
Well-researched responses based on relevant abstracts and paper content.
Custom Instructions Pro
Preferences or requirements that you'd like Emergent Mind to consider when generating responses.
Gemini 2.5 Flash
Gemini 2.5 Flash 87 tok/s
Gemini 2.5 Pro 51 tok/s Pro
GPT-5 Medium 17 tok/s Pro
GPT-5 High 23 tok/s Pro
GPT-4o 102 tok/s Pro
Kimi K2 166 tok/s Pro
GPT OSS 120B 436 tok/s Pro
Claude Sonnet 4 37 tok/s Pro
2000 character limit reached

Euclid Quick Data Release (Q1): The Strong Lensing Discovery Engine A -- System overview and lens catalogue (2503.15324v1)

Published 19 Mar 2025 in astro-ph.GA

Abstract: We present a catalogue of 497 galaxy-galaxy strong lenses in the Euclid Quick Release 1 data (63 deg$2$). In the initial 0.45\% of Euclid's surveys, we double the total number of known lens candidates with space-based imaging. Our catalogue includes 250 grade A candidates, the vast majority of which (243) were previously unpublished. Euclid's resolution reveals rare lens configurations of scientific value including double-source-plane lenses, edge-on lenses, complete Einstein rings, and quadruply-imaged lenses. We resolve lenses with small Einstein radii ($\theta_{\rm E} < 1''$) in large numbers for the first time. These lenses are found through an initial sweep by deep learning models, followed by Space Warps citizen scientist inspection, expert vetting, and system-by-system modelling. Our search approach scales straightforwardly to Euclid Data Release 1 and, without changes, would yield approximately 7000 high-confidence (grade A or B) lens candidates by late 2026. Further extrapolating to the complete Euclid Wide Survey implies a likely yield of over 100000 high-confidence candidates, transforming strong lensing science.

Summary

Euclid Quick Data Release and Strong Lensing Discovery

The paper presents an analysis of the Euclid Quick Data Release (Q1) focusing on the discovery of strong gravitational lenses utilizing state-of-the-art methodologies. The discovery is facilitated by the innovative "Strong Lensing Discovery Engine," which leverages a combination of deep learning, citizen science, expert validation, and sophisticated modeling techniques. With an integrated approach, the paper aims to manage the vast data within Euclid's surveys, enhancing the efficiency and accuracy of lens discovery.

The Euclid space telescope offers unparalleled capabilities in the field of strong lensing due to its space-based resolution and wide field-of-view, making it uniquely suitable for hunting galaxy-galaxy lens systems. The telescope's advantage over ground-based instruments typically lies in its superior point spread function (PSF), allowing the resolve of lenses with smaller Einstein radii that are often beyond the capabilities of terrestrial surveys.

Methodology Overview

  1. Deep Learning Models: Initially, an ensemble of deep learning models, including the Zoobot foundation model, was trained using simulated and real images. Zoobot emerged as the top performer, suggesting that models pre-trained on extensive real data exhibit superior generalization in the detection of strong lenses over those solely relying on simulations.
  2. Citizen Science Initiative: Images flagged by the neural networks were submitted to the Space Warps platform, which harnesses citizen scientists' efforts to sift through vast numbers of images and highlight likely lens candidates. Volunteers used machine learning-assisted scoring to refine the initial selections, improving the purity of the candidate set by excluding obvious non-lenses and potential misclassifications.
  3. Expert Validation: The next stage involved professional astronomers who evaluated the candidate lenses, assigning grades based on confidence levels with additional ratings for scientifically valuable configurations such as double source plane lenses. Their assessments guided further quantitative lens modeling.
  4. Automated Lens Modeling: Candidates were then subjected to automated modeling using PyAutoLens, which confirmed the genuineness of the lens by validating the inferred physical structures. This modeling extended beyond visual identification, estimating key parameters such as the Einstein radius and mass distribution.

Results and Implications

From Q1's 63 deg² field, the paper identified 250 high-confidence lensing systems previously unknown, effectively doubling the known space-based lens sample. The paper quantitatively confirms \num{7000} high-confidence lens candidates feasible by late 2026 for Euclid's Data Release 1 without modifications to the current discovery engine, with potential up to over \num{100000} candidates in the wide survey.

The implications of this increase in identified lens systems are substantial. The ability to resolve lenses down to very small Einstein radii boosts opportunities for a variety of scientific enquiries, including the precise measurement of dark matter distributions within galaxies and the characterization of galaxy evolution processes. It also opens pathways to more reliable cosmological constraints, particularly regarding dark energy models and gravitational lensing effects.

Future work would focus on refining machine-learning models' parameter sensitivity and integration of more sophisticated ensemble methods to optimize discovery, coupled with large-scale spectroscopic follow-up to characterize the physical properties of the identified lens systems. With Euclid poised for further releases, the methodology promises a transformative leap in the quantity and quality of gravitational lens discoveries.

Lightbulb Streamline Icon: https://streamlinehq.com

Continue Learning

We haven't generated follow-up questions for this paper yet.

List To Do Tasks Checklist Streamline Icon: https://streamlinehq.com

Collections

Sign up for free to add this paper to one or more collections.

X Twitter Logo Streamline Icon: https://streamlinehq.com

Tweets

This paper has been mentioned in 2 posts and received 133 likes.

Don't miss out on important new AI/ML research

See which papers are being discussed right now on X, Reddit, and more:

“Emergent Mind helps me see which AI papers have caught fire online.”

Philip

Philip

Creator, AI Explained on YouTube