The volume of data that will be produced by new-generation surveys requires automatic classification methods to select and analyze sources. Indeed, this is the case for the search for strong gravitational lenses, where the population of the detectable lensed sources is only a very small fraction of the full source population. We apply for the first time a morphological classification method based on a Convolutional Neural Network (CNN) for recognizing strong gravitational lenses in 255 square degrees of the Kilo Degree Survey (KiDS), one of the current-generation optical wide surveys. The CNN is currently optimized to recognize lenses with Einstein radii > ∼ 1.4 arcsec, about twice the r-band seeing in KiDS. In a sample of 21789 colour-magnitude selected Luminous Red Galaxies (LRG), of which three are known lenses, the CNN retrieves 761 strong-lens candidates and correctly classifies two out of three of the known lenses. The misclassified lens has an Einstein radius below the range on which the algorithm is trained. We down-select the most reliable 56 candidates by a joint visual inspection. This final sample is presented and discussed. A conservative estimate based on our results shows that with our proposed method it should be possible to find ∼ 100 massive LRG-galaxy lenses at z ∼ < 0.4 in KiDS when completed. In the most optimistic scenario this number can grow considerably (to maximally ∼2400 lenses), when widening the colour-magnitude selection and training the CNN to recognize smaller image-separation lens systems.

We present a new sample of galaxy-scale strong gravitational-lens candidates, selected from 904 square degrees of Data Release 4 of the Kilo-Degree Survey (KiDS), i.e., the "Lenses in the Kilo-Degree Survey" (LinKS) sample. We apply two Convolutional Neural Networks (ConvNets) to ∼ 88 000 colour-magnitude selected luminous red galaxies yielding a list of 3500 strong-lens candidates. This list is further down-selected via human inspection. The resulting LinKS sample is composed of 1983 rank-ordered targets classified as "potential lens candidates" by at least one inspector. Of these, a high-grade subsample of 89 targets is identified with potential strong lenses by all inspectors. Additionally, we present a collection of another 200 strong lens candidates discovered serendipitously from various previous ConvNet runs. A straightforward application of our procedure to future Euclid or LSST data can select a sample of ∼ 3000 lens candidates with less than 10 per cent expected false positives and requiring minimal human intervention.

Convolutional Neural Networks (ConvNets) are one of the most promising methods for identifying strong gravitational lens candidates in survey data. We present two ConvNet lens-finders which we have trained with a dataset composed of real galaxies from the Kilo Degree Survey (KiDS) and simulated lensed sources. One ConvNet is trained with single r -band galaxy images, hence basing the classification mostly on the morphology. While the other ConvNet is trained on g-r-i composite images, relying mostly on colours and morphology. We have tested the ConvNet lens-finders on a sample of 21789 Luminous Red Galaxies (LRGs) selected from KiDS and we have analyzed and compared the results with our previous ConvNet lens-finder on the same sample. The new lens-finders achieve a higher accuracy and completeness in identifying gravitational lens candidates, especially the single-band ConvNet. Our analysis indicates that this is mainly due to improved simulations of the lensed sources. In particular, the single-band ConvNet can select a sample of lens candidates with ∼ 40% purity, retrieving 3 out of 4 of the confirmed gravitational lenses in the LRG sample. With this particular setup and limited human intervention, it will be possible to retrieve, in future surveys such as Euclid, a sample of lenses exceeding in size the total number of currently known gravitational lenses.

We report new high-quality galaxy-scale strong lens candidates found in the Kilo-Degree Survey data release 4 using machine learning. We have developed a new convolutional neural network (CNN) classifier to search for gravitational arcs, following the prescription by Petrillo et al. and using only r-band images. We have applied the CNN to two "predictive samples": a luminous red galaxy (LRG) and a "bright galaxy" (BG) sample (r<21). We have found 286 new high-probability candidates, 133 from the LRG sample and 153 from the BG sample. We have ranked these candidates based on a value that combines the CNN likelihood of being a lens and the human score resulting from visual inspection (P-value), and here we present the highest 82 ranked candidates with P-values 0.5. All of these high-quality candidates have obvious arc or pointlike features around the central red defector. Moreover, we define the best 26 objects, all with P-values 0.7, as a "golden sample" of candidates. This sample is expected to contain very few false positives; thus, it is suitable for follow-up observations. The new lens candidates come partially from the more extended footprint adopted here with respect to the previous analyses and partially from a larger predictive sample (also including the BG sample). These results show that machine-learning tools are very promising for finding strong lenses in large surveys and more candidates can be found by enlarging the predictive samples beyond the standard assumption of LRGs. In the future, we plan to apply our CNN to the data from next-generation surveys such as the Large Synoptic Survey Telescope, Euclid, and the Chinese Space Station Optical Survey.

Constraining the sub-galactic matter-power spectrum on 1-10 kpc scales would make it possible to distinguish between the concordance ΛCDM model and various alternative dark-matter models due to the significantly different levels of predicted mass structure. Here, we demonstrate a novel approach to observationally constrain the population of overall law-mass density fluctuations in the inner regions of massive elliptical lens galaxies, based on the power spectrum of the associated surface-brightness perturbations observable in highly magnified galaxy-scale Einstein rings and gravitational arcs. The application of our method to the SLACS lens system SDSS J0252+0039 results in the following limits (at the 99 per cent confidence level) on the dimensionless convergence-power spectrum (and the associated standard deviation in aperture mass): ∆ 2 δκ < 1 (σ AM < 0.8 × 10 8 M ) on 0.5-kpc scale, ∆ 2 δκ < 0.1 (σ AM < 1 × 10 8 M ) on 1-kpc scale and ∆ 2 δκ < 0.01 (σ AM < 3 × 10 8 M ) on 3-kpc scale. The estimated effect of CDM sub-haloes lies considerably below these first observational upper-limit constraints on the level of inhomogeneities in the projected total mass distribution of galactic haloes. Future analysis for a larger sample of galaxy-galaxy strong lens systems will narrow down these constraints and rule out all cosmological models predicting a significantly larger level of clumpiness on these critical sub-galactic scales.