STRADAViT is a self‑supervised continued‑pretraining framework that adapts Vision Transformer (ViT) backbones for radio‑astronomy image analysis. It curates mixed‑survey data, generates radio‑astronomy‑aware training views, and initializes encoders with ViT‑MAE, optionally adding register tokens. Evaluations on three morphology benchmarks (MiraBest, LoTSS DR2, and Radio Galaxy Zoo) show that a register‑based two‑stage checkpoint improves linear‑probe Macro‑F1 scores over the ViT‑MAE baseline and enhances fine‑tuning on MiraBest and RGZ DR1, though performance on LoTSS DR2 fine‑tuning declines; these differences are statistically significant.
By Andrea DeMarco, Ian Fenech Conti, Hayley Camilleri, Ardiana Bushi, Simone Riggi
The study audits the AION-1 foundation model, a 39‑modality transformer trained on over 200 million astronomical objects, and finds that its reliance on a survey detection channel—specifically the segmentation map—introduces a severe systematic bias. By keeping image tokens unchanged and editing only the segmentation map, all model outputs (flux, size, ellipticity, redshift) shift by factors of 110–4400 compared to a placebo, revealing that the model’s predictions are driven more by detection gating than by the actual light distribution. This bias propagates into cosmological analyses, shifting tomographic mean redshifts by a median 0.71 × the LSST DESC requirement and exceeding it in multiple assignments, while removing the detection channel eliminates the effect without measurable cost.
whyItMatters":"The bias in the detection channel directly inflates errors in key astronomical measurements, potentially compromising the precision of cosmological studies that rely on accurate redshift estimates."
By Ihor Kendiukhov
The paper presents a probabilistic machine‑learning framework for classifying low‑ and high‑resolution stellar and extragalactic spectra, targeting the upcoming 4MOST survey. Four approaches were evaluated—CNNs, Dirichlet distribution, Monte Carlo dropout (MCD), and Bayesian neural networks with variational inference—using SDSS data and a 4MOST mock dataset. The MCD‑augmented CNN achieved the highest accuracies (92.6% on SDSS, 93.9% on mock data) while also delivering well‑calibrated uncertainty estimates with minimal extra computational cost.
By Simon Barton, Martin Sahl\'en, Andreas Korn, Christian Glaser
arXiv:2609.35966v1 Announce Type: cross
Abstract: Radio-interferometric images are reconstructed from sparsely sampled visibilities, and CLEAN-based imaging can struggle with spatial filtering, compl...
By Shunyuan Mao, Andrea Isella, Paris Perdikaris, Li-Ta Lo, Hui Li
We compare feature-based machine learning and end-to-end deep learning for variable star classification using the Zwicky Transient Facility (ZTF) StarEmbed dataset, which contains about 40,000 objects...
The study adapts a convolutional neural network to classify galaxy morphologies using crowd-sourced annotations from Galaxy Zoo 1. It evaluates how training strategies—such as training all layers versus only the last, incorporating hierarchical labels, varying data volume and annotator agreement, staged transfer learning, and ensembling—affect accuracy and efficiency. Results show that full-network training and high annotator agreement yield over 99% accuracy, while hierarchical approaches and staged learning help when data are limited.
By Luis Enrique Sucar, Carlos del Burgo, Jonathan Serrano-P\'erez
arXiv:2606. 05103v1 Announce Type: new Abstract: The Nancy Grace Roman Space Telescope (Roman), set for launch as early as September 2026, will conduct wide-field infrared imaging surveys with unprecedented spatial resolution and cadence, enabling the discovery of millions of astronomical transients.
By Karan Gandhi, Ashish A. Mahabal, Jacob E. Jencson, Russ R. Laher, Ben Rusholme, Lin Yan, Ryan M. Lau, Schuyler D. Van Dyk, Mansi M. Kasliwal
arXiv:2606. 02778v1 Announce Type: cross Abstract: I present EXOVEIL, a transit detection system that learns what a star's brightness should look like and flags when reality disagrees.
By Pratik Priyanshu
The paper presents a two‑stage approach to street‑panorama semantic segmentation using Vision Transformer ensembles. In the first stage, nine pretrained models are compared under equal compute budgets, and the best two are trained with three random seeds, achieving 60.95% mIoU on validation and 57.08% on the hidden test leaderboard. The second stage explores augmentations, larger inputs, and additional architectures, culminating in an ensemble of four models that scores 64.53% mIoU on validation and 61.54% mIoU (71.93% F1) on the hidden test leaderboard, while the study documents all runs, costs, and releases the code and models.
By Yunus Serhat B{\i}\c{c}ak\c{c}{\i}
The study presents a deep‑learning approach (YOLOv11) trained on SDSS images to distinguish genuine dual active galactic nuclei (DAGN) from chance superpositions and foreground stars in the GOTHIC survey. Applying the model to 46,061 previously rejected candidates yields 29,605 dual‑nucleus candidates, with 54.5–62 % likely genuine, and a conservative subset of ~13,672 compact systems. The resulting catalogue refines the DAGN candidate list, reducing contamination and expanding the plausible census, though spectroscopic confirmation remains needed.
By Bhavesh Mukheja, Snehanshu Saha, Anwesh Bhattacharya, Mousumi Das, Fran\c{c}oise Combes, Sudhanshu Barway
arXiv:2606. 07882v1 Announce Type: cross Abstract: Different vision neural networks -- trained to classify, contrast, reconstruct, or match images to text -- should have correspondingly different internal representations.
By Yousef Radwan
The study evaluates EfficientNet variants B0–B7 for classifying retinal diseases from spectral-domain optical coherence tomography (SD‑OCT) images at a fixed 224×224 resolution. Eight ImageNet‑pretrained models were tested on the Kermany dataset, with training involving a frozen‑backbone head followed by partial fine‑tuning. Results show non‑monotonic accuracy, with B1 achieving the highest overall accuracy (96.95%) and best Drusen F1, while B7 offered a secondary accuracy gain at the cost of larger model size and CPU latency.
By Shkelqim Sherifi