Accurate classification of celestial objects is essential for advancing our understanding of the universe. MargNet is a recently developed deep learning-based classifier applied to SDSS DR16 dataset to segregate stars, quasars, and compact galaxies using photometric data. MargNet utilizes a stacked architecture, combining a Convolutional Neural Network (CNN) for image modelling and an Artificial Neural Network (ANN) for modelling photometric parameters. In this study, we propose enhancing MargNet's performance by incorporating attention mechanisms and Vision Transformer (ViT)-based models for processing image data. The attention mechanism allows the model to focus on relevant features and capture intricate patterns within images, effectively distinguishing between different classes of celestial objects. Additionally, we leverage ViTs, a transformer-based deep learning architecture renowned for exceptional performance in image classification tasks. We enhance the model's understanding of complex astronomical images by utilizing ViT's ability to capture global dependencies and contextual information. Our approach uses a curated dataset comprising 240,000 compact and 150,000 faint objects. The models learn classification directly from the data, minimizing human intervention. Furthermore, we explore ViT as a hybrid architecture that uses photometric features and images together as input to predict astronomical objects. Our results demonstrate that the proposed attention mechanism augmented CNN in MargNet marginally outperforms the traditional MargNet and the proposed ViT-based MargNet models. Additionally, the ViT-based hybrid model emerges as the most lightweight and easy-to-train model with classification accuracy similar to that of the best-performing attention-enhanced MargNet.
Archives of Photographic PLates for Astronomical USE (APPLAUSE) Digitisation of astronomical plates and their integration into the International Virtual Observatory
AstronomicAL is a human-in-the-loop interactive labelling and training dashboard that allows users to create reliable datasets and robust classifiers using active learning. This technique prioritises data that offer high information gain, leading to improved performance using substantially less data. The system allows users to visualise and integrate data from different sources and deal with incorrect or missing labels and imbalanced class sizes. AstronomicAL enables experts to visualise domain-specific plots and key information relating both to broader context and details of a point of intere…
Applying Vision Transformers on Spectral Analysis of Astronomical Objects
Harvard Extension School（哈佛大学延伸学校）；Harvard University（哈佛大学）；John A. Paulson School of Engineering and Applied Science（约翰·A·保罗森工程与应用科学学院）；Department of Computer Science（计算机科学系）；Center for Data and Artificial Intelligence（数据与人工智能中心）；Millennium Institute of Astrophysics (MAS)（千年天体物理研究所 (MAS)）；Millennium Nucleus for Galaxies (MINGAL)（银河千年核 (MINGAL)）；Heidelberg Institute for Theoretical Studies（海德堡理论研究所）
Detection, Instance Segmentation, and Classification for Astronomical Surveys with Deep Learning (DeepDISC): Detectron2 Implementation and Demonstration with Hyper Suprime-Cam Data
The Virtual Observatory (VO) is realizing global electronic integration of astronomy data. One of the long-term goals of the U.S. VO project, the Virtual Astronomical Observatory (VAO), is development of services and protocols that respond to the growing size and complexity of astronomy data sets. This paper describes how VAO staff are active in such development efforts, especially in innovative strategies and techniques that recognize the limited operating budgets likely available to astronomers even as demand increases. The project has a program of professional outreach whereby new services …
University College London（伦敦大学学院）；Université Paris-Saclay（巴黎萨克雷大学）；The Alan Turing Institute（艾伦·图灵研究所）
A new golden age in astronomy is upon us, dominated by data. Large astronomical surveys are broadcasting unprecedented rates of information, demanding machine learning as a critical component in modern scientific pipelines to handle the deluge of data. The upcoming Legacy Survey of Space and Time (LSST) of the Vera C. Rubin Observatory will raise the big-data bar for time-domain astronomy, with an expected 10 million alerts per-night, and generating many petabytes of data over the lifetime of the survey. Fast and efficient classification algorithms that can operate in real-time, yet robustly a…
We demonstrate that extremely rapid and weak periodic and non-periodic signals can easily be detected by using the autocorrelation of intensity as a function of time. We use standard radio-astronomical observations that have artificial periodic and non-periodic signals generated by the electronics of terrestrial origin. The autocorrelation detects weak signals that have small amplitudes because it averages over long integration times. Another advantage is that it allows a direct visualization of the shape of the signals, while it is difficult to see the shape with a Fourier transform. Although…
I present a software tool for solving the astrometry of astronomical images. The code puts emphasis on robustness against failures for correctly matching the sources in the image to a reference catalog, and on the stability of the solutions over the field of view (e.g., using orthogonal polynomials for the fitted transformation). The code was tested on over 50,000 images from various sources, including the Palomar Transient Factory (PTF) and the Zwicky Transient Facility (ZTF). The tested images equally represent low and high Galactic latitude fields and exhibit failure/bad-solution rate of <2…
Finding and classifying astronomical sources is key in the scientific exploitation of radio surveys. Source-finding usually involves identifying the parts of an image belonging to an astronomical source, against some estimated background. This can be problematic in the radio regime, owing to the presence of correlated noise, which can interfere with the signal from the source. In the current work, we present ConvoSource, a novel method based on a deep learning technique, to identify the positions of radio sources, and compare the results to a Gaussian-fitting method. Since the deep learning ap…
Enhancing astrometric registration of Chinese historical Astronomical Digital Plates with deep learning
Shanghai Astronomical Observatory, Chinese Academy of Sciences（中国科学院上海天文台）；School of Astronomy and Space Science, University of Chinese Academy of Sciences（中国科学院大学天文与空间科学学院）
We have designed and implemented a novel way to process wide-field astronomical data within a distributed environment of hardware resources and humanpower. The system is characterized by integration of archiving, calibration, and post-calibration analysis of data from raw, through intermediate, to final data products. It is a true integration thanks to complete linking of data lineage from the final catalogs back to the raw data. This paper describes the pipeline processing of optical wide-field astronomical data from the WFI (http://www.eso.org/lasilla/instruments/wfi/) and OmegaCAM (http://w…
The Montage Image Mosaic Engine was designed as a scalable toolkit, written in C for performance and portability across *nix platforms, that assembles FITS images into mosaics. The code is freely available and has been widely used in the astronomy and IT communities for research, product generation and for developing next-generation cyber-infrastructure. Recently, it has begun to finding applicability in the field of visualization. This has come about because the toolkit design allows easy integration into scalable systems that process data for subsequent visualization in a browser or client. …
The U.S. Virtual Astronomical Observatory was a software infrastructure and development project designed both to begin the establishment of an operational Virtual Observatory (VO) and to provide the U.S. coordination with the international VO effort. The concept of the VO is to provide the means by which an astronomer is able to discover, access, and process data seamlessly, regardless of its physical location. This paper describes the origins of the VAO, including the predecessor efforts within the U.S. National Virtual Observatory, and summarizes its main accomplishments. These accomplishmen…
We present an overview of best practices for publishing data in astronomy and astrophysics journals. These recommendations are intended as a reference for authors to help prepare and publish data in a way that will better represent and support science results, enable better data sharing, improve reproducibility, and enhance the reusability of data. Observance of these guidelines will also help to streamline the extraction, preservation, integration and cross-linking of valuable data from astrophysics literature into major astronomical databases, and consequently facilitate new modes of science…
Forthcoming instruments designed for high-cadence large-area surveys, such as the Dark Energy Survey and Large Synoptic Survey Telescope, will generate several GB of data products every few minutes during survey operations. Since such surveys are designed to operate with minimal observer interaction, automated real-time analysis of these large images is necessary to ensure uninterrupted production of science-quality data. We describe a software infrastructure suite designed to support such surveys, focusing particularly on ImageHealth, a tool for near-real-time processing of large images. Thes…
CASA, the Common Astronomy Software Applications, is the primary data processing software for the Atacama Large Millimeter/submillimeter Array (ALMA) and the Karl G. Jansky Very Large Array (VLA), and is often used also for other radio telescopes. CASA has always been distributed as a single, integrated application, including a Python interpreter and all the libraries, packages and modules. As part of the ongoing development of CASA 6, and the switch from Python 2 to 3, CASA will provide greater flexibility for users to integrate CASA into existing Python workflows by using a modular architect…
Transformer-Based Neural Network for Transient Detection without Image Subtraction
Department of Physics and Astronomy, University of Pennsylvania（物理与天文学系，宾夕法尼亚大学）；Department of Physics and Astronomy, University of Delaware（物理与天文学系，德克萨斯大学）；University of Delaware, Joseph R. Biden, Jr. School of Public Policy and Administration（德克萨斯大学乔伊斯·拜登初级公共政策与行政学校）；University of Delaware, Data Science Institute（德克萨斯大学数据科学研究所）
Astrophysical light curves are particularly challenging data objects due to the intensity and variety of noise contaminating them. Yet, despite the astronomical volumes of light curves available, the majority of algorithms used to process them are still operating on a per-sample basis. To remedy this, we propose a simple Transformer model -- called Denoising Time Series Transformer (DTST) -- and show that it excels at removing the noise and outliers in datasets of time series when trained with a masked objective, even when no clean targets are available. Moreover, the use of self-attention ena…
Innovative developments in data processing, archiving, analysis, and visualization are nowadays unavoidable to deal with the data deluge expected in next-generation facilities for radio astronomy, such as the Square Kilometre Array (SKA) and its precursors. In this context, the integration of source extraction and analysis algorithms into data visualization tools could significantly improve and speed up the cataloguing process of large area surveys, boosting astronomer productivity and shortening publication time. To this aim, we are developing a visual analytic platform (CIRASA) for advanced …
Astronomical image time series classification using CONVolutional attENTION (ConvEntion)
University of Montpellier（蒙彼利埃大学）；University of Nimes（尼姆大学）；Aix Marseille Univ（艾克斯-马赛大学）；Paul Valéry University Montpellier 3（蒙彼利埃第三大学（保罗·瓦莱里大学））；INRAE（法国国家农业、食品与环境研究院）；CIRAD（法国国际农业研究合作中心）；CNRS（法国国家科学研究中心）
StarEmbed: Benchmarking Time Series Foundation Models on Astronomical Observations of Variable Stars
Department of Computer Science, Northwestern University（计算机科学系，西北大学）；Center for Foundation Models and Generative AI, Northwestern University（基础模型与生成AI中心，西北大学）；NSF – Simons AI Institute for the Sky (SkAI)（国家科学基金会-斯隆人工智能天文研究所（SkAI））；Department of Physics and Astronomy, Northwestern University（物理与天文学系，西北大学）；Center for Interdisciplinary Exploration and Research in Astrophysics (CIERA)（天文学跨学科探索与研究中心（CIERA））；Department of Statistics and Data Science, Northwestern University（统计与数据科学系，西北大学）
I describe a new, open-source astronomical image-fitting program called Imfit, specialized for galaxies but potentially useful for other sources, which is fast, flexible, and highly extensible. A key characteristic of the program is an object-oriented design which allows new types of image components (2D surface-brightness functions) to be easily written and added to the program. Image functions provided with Imfit include the usual suspects for galaxy decompositions (Sersic, exponential, Gaussian), along with Core-Sersic and broken-exponential profiles, elliptical rings, and three components …
Astronomical data are gathered through a very large number of heterogeneous techniques and stored in very diversified and often incompatible data repositories. Moreover in the e-science environment, it is needed to integrate services across distributed, heterogeneous, dynamic "virtual organizations" formed by different resources within a single enterprise and/or external resource sharing and service provider relationships. The DAME/VONeural project, run jointly by the University Federico II, INAF (National Institute of Astrophysics) Astronomical Observatories of Napoli and the California Insti…
A High-Throughput AES-GCM Implementation on GPUs for Secure, Policy-Based Access to Massive Astronomical Catalogs
Light Bridges S.L.（Light Bridges 有限公司）；Observatorio Astronómico del Teide（泰德天文台）；Universidad de La Laguna (ULL)（拉古纳大学）；Instituto Tecnológico y de Energías Renovables (ITER)（技术与可再生能源研究所）；Instituto de Astrofísica de Canarias (IAC)（加那利天体物理研究所）
Deeper detection limits in astronomical imaging using self-supervised spatiotemporal denoising
Department of Automation, Beijing National Research Center for Information Science and Technology, Tsinghua University, Beijing, China（自动化系，北京信息科学与技术国家研究中心，清华大学，北京，中国）；Department of Astronomy, Tsinghua University, Beijing, China（天文系，清华大学，北京，中国）；Institute for Brain and Cognitive Sciences, Tsinghua University, Beijing, China（脑与认知科学研究院，清华大学，北京，中国）；Kavli Institute for the Physics and Mathematics of the Universe (WPI), The University of Tokyo, Kashiwa, Japan（宇宙物理与数学科学研究所（WPI），东京大学，日光，日本）；National Astronomical Observatory of Japan, Tokyo, Japan（日本国家天文台，东京，日本）；Institute for Cosmic Ray Research, The University of Tokyo, Kashiwa, Japan（宇宙射线研究所，东京大学，日光，日本）；College of AI, Tsinghua University, Beijing, China（人工智能学院，清华大学，北京，中国）；School of Mathematics and Physics, Qinghai University, Xining, China（数学与物理学院，青海大学，西宁，中国）；Tsinghua Wuxi Institute of Applied Technology, Wuxi, China（清华大学无锡应用技术研究院，无锡，中国）
XAMI -- A Benchmark Dataset for Artefact Detection in XMM-Newton Optical Images
Department of Computers and Information Technology, Politehnica University of Timişoara（蒂米șoara 技术大学计算机与信息科技系）；European Space Agency (ESA)（欧洲航天局）；European Space Astronomy Centre (ESAC)（欧洲空间天文台）；Serco Ltd.（Serco 公司）
iDaVIE v1.0: A virtual reality tool for interactive analysis of astronomical data cubes
Inter-University Institute for Data Intensive Astronomy, University of Cape Town（开普敦大学数据密集型天文学跨大学研究所）；Department of Astronomy, University of Cape Town（开普敦大学天文系）；Armagh Observatory and Planetarium（阿马赫天文台和天文馆）；INAF - Istituto di Radioastronomia（意大利国家天体物理研究所射电天文研究所）；Astrophysical Observatory of Catania, National Institute for Astrophysics（卡塔尼亚天体物理观测站，意大利国家天体物理研究所）；Kapteyn Astronomical Institute, University of Groningen（格罗宁根大学卡普坦天文研究所）
