Reducing false positives in strong lens detection through effective augmentation and ensemble learning

Abstract

This research studies the impact of high-quality training datasets on the performance of Convolutional Neural Networks (CNNs) in detecting strong gravitational lenses. We stress the importance of data diversity and representativeness, demonstrating how variations in sample populations influence CNN performance. In addition to the quality of training data, our results highlight the effectiveness of various techniques, such as data augmentation and ensemble learning, in reducing false positives while maintaining model completeness at an acceptable level. This enhances the robustness of gravitational lens detection models and advancing capabilities in this field. Our experiments, employing variations of DenseNet and EfficientNet, achieved a best false positive rate (FP rate) of $10^{-4}$, while successfully identifying over 88 per cent of genuine gravitational lenses in the test dataset. This represents an 11-fold reduction in the FP rate compared to the original training dataset. Notably, this substantial enhancement in the FP rate is accompanied by only a 2.3 per cent decrease in the number of true positive samples. Validated on the KiDS dataset, our findings offer insights applicable to ongoing missions, like Euclid.

15 pages, 14 figures, 7 tables, Accepted for publication in MNRAS

References in corpus (56)

• H0LiCOW XIII. A 2.4% measurement of $H_0$ from lensed quasars: $5.3\sigma$ tension between early and late-Universe probes
• Galaxy Zoo 2: detailed morphological classifications for 304,122 galaxies from the Sloan Digital Sky Survey
• The Sloan Lens ACS Survey. X. Stellar, Dynamical, and Total Mass Correlations of Massive Early-type Galaxies
• The Sloan Lens ACS Survey. III - The Structure and Formation of Early-type Galaxies and their Evolution since z~1
• H0LiCOW V. New COSMOGRAIL time delays of HE0435-1223: $H_0$ to 3.8% precision from strong lensing in a flat $\Lambda$CDM model
• Strong Lensing by Galaxies
• The Sloan Lens ACS Survey. V. The Full ACS Strong-Lens Sample
• Dissecting the Gravitational Lens B1608+656. II. Precision Measurements of the Hubble Constant, Spatial Curvature, and the Dark Energy Equation of State
• Gravitational detection of a low-mass dark satellite at cosmological distance
• Identifying Exoplanets with Deep Learning: A Five Planet Resonant Chain around Kepler-80 and an Eighth Planet around Kepler-90
• Two accurate time-delay distances from strong lensing: Implications for cosmology
• The fourth data release of the Kilo-Degree Survey: ugri imaging and nine-band optical-IR photometry over 1000 square degrees
• The Sloan Lens ACS Survey. IX. Colors, Lensing and Stellar Masses of Early-type Galaxies
• Warm dark matter chills out: constraints on the halo mass function and the free-streaming length of dark matter with 8 quadruple-image strong gravitational lenses
• The Sloan-Lens ACS Survey II: stellar populations and internal structure of early-type lens galaxies
• Inference of the Cold Dark Matter substructure mass function at z=0.2 using strong gravitational lenses
• CMU DeepLens: Deep Learning For Automatic Image-based Galaxy-Galaxy Strong Lens Finding
• Evidence For A Mild Steepening And Bottom-Heavy IMF In Massive Galaxies From Sodium And Titanium-Oxide Indicators
• Finding Strong Gravitational Lenses in the Kilo Degree Survey with Convolutional Neural Networks
• SHARP -- VII. New constraints on the dark matter free-streaming properties and substructure abundance from gravitationally lensed quasars
• HerMES: Candidate Gravitationally Lensed Galaxies and Lensing Statistics at Submillimeter Wavelengths
• The stellar IMF determined in early-type galaxies from a non-degenerate set of optical line indices
• Cosmology with Strong Lensing Systems
• Constraints on Cosmological Parameters from the Analysis of the Cosmic Lens All Sky Survey Radio-Selected Gravitational Lens Statistics
• The Strong Gravitational Lens Finding Challenge
• Galaxy Zoo: An independent look at the evolution of the bar fraction over the last eight billion years from HST-COSMOS
• Convolutional neural networks: a magic bullet for gravitational-wave detection?
• Optimization of the Observing Cadence for the Rubin Observatory Legacy Survey of Space and Time: a pioneering process of community-focused experimental design
• Classifying Radio Galaxies with Convolutional Neural Network
• Low-mass halo perturbations in strong gravitational lenses at redshift z$\sim$0.5 are consistent with CDM
• Space Warps: I. Crowd-sourcing the Discovery of Gravitational Lenses
• Space Warps II. New Gravitational Lens Candidates from the CFHTLS Discovered through Citizen Science
• Recent Advances in Deep Learning: An Overview
• LinKS: Discovering galaxy-scale strong lenses in the Kilo-Degree Survey using Convolutional Neural Networks
• The image separation distribution of strong lenses: Halo versus subhalo populations
• TDCOSMO V: strategies for precise and accurate measurements of the Hubble constant with strong lensing
• Models for the lens and source of B0218+357 - A LensClean approach to determine H0
• New high-quality strong lens candidates with deep learning in the Kilo Degree Survey
• Properties and characteristics of the WFIRST H4RG-10 detectors
• Gravitational lenses and lens candidates identified from the COSMOS field
• High-quality strong lens candidates in the final Kilo Degree survey footprint
• SHARP - V. Modelling gravitationally-lensed radio arcs imaged with global VLBI observations
• CNN Architecture Comparison for Radio Galaxy Classification
• Testing Convolutional Neural Networks for finding strong gravitational lenses in KiDS
• Convolutional Neural Networks on the HEALPix sphere: a pixel-based algorithm and its application to CMB data analysis
• Convolutional neural networks for the detection of the early inspiral of a gravitational-wave signal
• Deep Learning for Strong Lensing Search: Tests of the Convolutional Neural Networks and New Candidates from KiDS DR3
• Testing dark energy models with a new sample of strong-lensing systems
• Convolutional Neural Networks for Direct Detection of Dark Matter
• Strong lensing statistics in large, z~<0.2 surveys: bias in the lens galaxy population
• A novel search for gravitationally lensed radio sources in wide-field VLBI imaging from the mJIVE-20 survey
• A machine learning based approach to gravitational lens identification with the International LOFAR Telescope
• Streamlined Lensed Quasar Identification in Multiband Images via Ensemble Networks
• Joint galaxy-galaxy lensing and clustering constraints on galaxy formation
• Galaxy Morphology Classification using EfficientNet Architectures
• Clusternets: A deep learning approach to probe clustering dark energy

Cited by in corpus (1)

• Assessing Galaxy Rotation Kinematics: Insights from Convolutional Neural Networks on Velocity Variations