Astronomy and Computing最新文献

{"title":"AstroMLab 1: Who wins astronomy jeopardy!?","authors":"Y.-S. Ting (丁源森) ,&nbsp;T.D. Nguyen ,&nbsp;T. Ghosal ,&nbsp;R. Pan (潘瑞) ,&nbsp;H. Arora ,&nbsp;Z. Sun (孙泽昌) ,&nbsp;T. de Haan ,&nbsp;N. Ramachandra ,&nbsp;A. Wells ,&nbsp;S. Madireddy ,&nbsp;A. Accomazzi","doi":"10.1016/j.ascom.2024.100893","DOIUrl":"10.1016/j.ascom.2024.100893","url":null,"abstract":"<div><div>We present a comprehensive evaluation of proprietary and open-weights large language models using the first astronomy-specific benchmarking dataset. This dataset comprises 4,425 multiple-choice questions curated from the Annual Review of Astronomy and Astrophysics, covering a broad range of astrophysical topics.<span><span>¹</span></span> Our analysis examines model performance across various astronomical subfields and assesses response calibration, crucial for potential deployment in research environments. Claude-3.5-Sonnet outperforms competitors by up to 4.6 percentage points, achieving 85.0% accuracy. For proprietary models, we observed a universal reduction in cost every 3-to-12 months to achieve similar score in this particular astronomy benchmark. open-weights models have rapidly improved, with LLaMA-3-70b (80.6%) and Qwen-2-72b (77.7%) now competing with some of the best proprietary models. We identify performance variations across topics, with non-English-focused models generally struggling more in exoplanet-related fields, stellar astrophysics, and instrumentation related questions. These challenges likely stem from less abundant training data, limited historical context, and rapid recent developments in these areas. This pattern is observed across both open-weights and proprietary models, with regional dependencies evident, highlighting the impact of training data diversity on model performance in specialized scientific domains. Top-performing models demonstrate well-calibrated confidence, with correlations above 0.9 between confidence and correctness, though they tend to be slightly underconfident. The development for fast, low-cost inference of open-weights models presents new opportunities for affordable deployment in astronomy. The rapid progress observed suggests that LLM-driven research in astronomy may become feasible in the near future.</div></div>","PeriodicalId":48757,"journal":{"name":"Astronomy and Computing","volume":"51 ","pages":"Article 100893"},"PeriodicalIF":1.9,"publicationDate":"2024-11-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142745130","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Extended black hole solutions in Rastall theory of gravity","authors":"M. Sharif ,&nbsp;M. Sallah","doi":"10.1016/j.ascom.2024.100897","DOIUrl":"10.1016/j.ascom.2024.100897","url":null,"abstract":"<div><div>We utilize the gravitational decoupling via the extended geometric deformation to extend the Schwarzschild vacuum solution to new black holes in Rastall theory. By employing linear transformations that deform both the temporal and radial coefficients of the metric, the field equations with a dual matter source are successfully decoupled into two sets. The first of these sets is described by the metric for the vacuum Schwarzschild spacetime, while the second set corresponds to the added extra source. Three extended solutions are obtained using two restrictions on the metric potentials and extra source, respectively. For selected values of the Rastall and decoupling parameters, we study the impact of the fluctuation of these parameters on the obtained models. We also investigate the asymptotic flatness of the resulting spacetimes by analysis of the metric coefficients. Finally, the nature of the additional source is explored for each model, via analysis of the energy conditions. It is found among other results that none of the obtained models satisfy the energy conditions, while only the model corresponding to the barotropic equation of state mimics an asymptotically flat spacetime.</div></div>","PeriodicalId":48757,"journal":{"name":"Astronomy and Computing","volume":"50 ","pages":"Article 100897"},"PeriodicalIF":1.9,"publicationDate":"2024-11-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142702064","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Classification of galaxies from image features using best parameter selection by horse herd optimization algorithm (HOA)","authors":"Ahmadreza Yeganehmehr,&nbsp;Hossein Ebrahimnezhad","doi":"10.1016/j.ascom.2024.100898","DOIUrl":"10.1016/j.ascom.2024.100898","url":null,"abstract":"<div><div>With the advancement of observation technology, visual data has made significant progress, rendering manual image classification less effective. Consequently, various image processing and automatic classification methods have garnered attention from researchers. Scientists estimate that there are approximately 2 trillion observable galaxies in the universe. Each galaxy possesses unique characteristics that are distinguishable. Therefore, finding a method to quickly and accurately identify these characteristics of each galaxy and classify them rapidly can greatly enhance the galaxy detection and classification process, while minimizing human errors. The objective of the present study is to determine the class of galaxies with from telescope image features using an optimized classifier with best parameters. The proposed method uses the HOA algorithm, based on the behavior of horse herds, to find the best parameters. This method evaluates the model's error with different SVM parameters and selects the optimal SVM parameters for constructing M-SVM. Using this method, the proposed algorithm is trained and ultimately applied to classify the test data. The results indicate that the developed model correctly classified up to 94.11% of the test dataset (1) and 90.74% of the test dataset (2).</div></div>","PeriodicalId":48757,"journal":{"name":"Astronomy and Computing","volume":"50 ","pages":"Article 100898"},"PeriodicalIF":1.9,"publicationDate":"2024-11-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142702063","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Accelerating radio astronomy imaging with RICK","authors":"E. De Rubeis ,&nbsp;G. Lacopo ,&nbsp;C. Gheller ,&nbsp;L. Tornatore ,&nbsp;G. Taffoni","doi":"10.1016/j.ascom.2024.100895","DOIUrl":"10.1016/j.ascom.2024.100895","url":null,"abstract":"<div><div>This paper presents an implementation of radio astronomy imaging algorithms on modern High Performance Computing (HPC) infrastructures, exploiting distributed memory parallelism and acceleration throughout multiple GPUs. Our code, called RICK (Radio Imaging Code Kernels), is capable of performing the major steps of the <span><math><mi>w</mi></math></span>-stacking algorithm presented in Offringa et al. (2014) both inter- and intra-node, and in particular has the possibility to run entirely on the GPU memory, minimising the number of data transfers between CPU and GPU. This feature, especially among multiple GPUs, is critical given the huge sizes of radio datasets involved.</div><div>After a detailed description of the new implementations of the code with respect to the first version presented in Gheller et al. (2023), we analyse the performances of the code for each step involved in its execution. We also discuss the pros and cons related to an accelerated approach to this problem and its impact on the overall behaviour of the code. Such approach to the problem results in a significant improvement in terms of runtime with respect to the CPU version of the code, as long as the amount of computational resources does not exceed the one requested by the size of the problem: the code, in fact, is now limited by the communication costs, with the computation that gets heavily reduced by the capabilities of the accelerators.</div></div>","PeriodicalId":48757,"journal":{"name":"Astronomy and Computing","volume":"50 ","pages":"Article 100895"},"PeriodicalIF":1.9,"publicationDate":"2024-11-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142702123","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A numerical solution of Schrödinger equation for the dynamics of early universe","authors":"M.Z. Mughal ,&nbsp;F. Khan","doi":"10.1016/j.ascom.2024.100894","DOIUrl":"10.1016/j.ascom.2024.100894","url":null,"abstract":"<div><div>Artificial neural networks (ANNs) have attained widespread success across varied disciplines. This study is designated for looking into an application of an integrated intelligent computing paradigm concerning dynamics in the early Universe through numerical solutions to the Schrödinger equation. To arrive at this we leverage the Levenberg–Marquardt backpropagation networks (LMBNs) to probe cosmic evolution in the early Universe with the Friedmann–Lemaitre–Robertson–Walker (FLRW) metric for a flat minisuperspace model of the Universe in the background. This leads to bridging quantum mechanics and inflationary Universe dynamics conducing to quantum cosmology within the standard model. Wheeler–DeWitt equation corresponds to the time-independent Schrödinger equation obtained from the equations of motion for a single scalar field in flat spacetime with FLRW metric. Utilizing the ntstool the whole computing process is operated for simulation. To evaluate the accuracy and efficiency of the proposed scheme a comparative analysis is carried out. To construct continuous neural network mappings we employ the explicit Runge–Kutta method as the target parameter for generating datasets. To determine the solution datasets of different scenarios, the training, testing, and validation processes are employed to take advantage of these in the learning of neural network models established upon the backpropagation technique of Levenberg–Marquardt. By varying related parameters we develop three scenarios that produce nine cases, three for each. The data plots of performance, training state, error histogram, regression, time-series response, and error autocorrelation represent the visualization of the results. These plots show a complete case description by displaying all the necessary data values. The analysis of these plots is presented to validate all the cases. Performing the analysis by mean square error (MSE) validates the achieved accuracy of the results by validating and verifying neural networks. This work is motivated by the compelling need to develop innovative computational methods for solving complex cosmological questions to untangle the conundrums of the early universe. The attractive numerical solutions of the Schrödinger equation for the early Universe heralds a step towards quantum cosmology based on the interplay of the Wheeler–DeWitt equation and time-independent Schrödinger equation. There is an increasing trend to use computational methods to solve ordinary and partial differential equations with the help of code development in Matlab. For this purpose feed-forward artificial neural network is used for investigating the Schrödinger equation.</div></div>","PeriodicalId":48757,"journal":{"name":"Astronomy and Computing","volume":"50 ","pages":"Article 100894"},"PeriodicalIF":1.9,"publicationDate":"2024-11-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142702126","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"The VSPEC Collection: A suite of utilities to model spectroscopic phase curves of 3D exoplanet atmospheres in the presence of stellar variability","authors":"Ted M. Johnson ,&nbsp;Cameron Kelahan ,&nbsp;Avi M. Mandell ,&nbsp;Ashraf Dhahbi ,&nbsp;Tobi Hammond ,&nbsp;Thomas Barclay ,&nbsp;Veselin B. Kostov ,&nbsp;Geronimo L. Villanueva","doi":"10.1016/j.ascom.2024.100890","DOIUrl":"10.1016/j.ascom.2024.100890","url":null,"abstract":"<div><div>We present the Variable Star PhasE Curve (<span>VSPEC</span>) Collection, a set of Python packages for simulating combined-light spectroscopic observations of 3-dimensional exoplanet atmospheres in the presence of stellar variability and inhomogeneity. <span>VSPEC</span> uses the Planetary Spectrum Generator’s Global Emission Spectra (PSG/GlobES) application along with a custom-built multi-component time-variable stellar model based on a user-defined grid of stellar photosphere models to produce spectroscopic light curves of the planet-host system. <span>VSPEC</span> can be a useful tool for modeling observations of exoplanets in transiting geometries (primary transit, secondary eclipse) as well as orbital phase curve measurements, and is built in a modular and flexible configuration for easy adaptability to new stellar and planetary model inputs. We additionally present a set of codes developed alongside the core <span>VSPEC</span> modules, including the stellar surface model generator <span>vspec-vsm</span>, the stellar spectral grid interpolation code GridPolator, and a Python interface for PSG, <span>libpypsg</span>.</div></div>","PeriodicalId":48757,"journal":{"name":"Astronomy and Computing","volume":"50 ","pages":"Article 100890"},"PeriodicalIF":1.9,"publicationDate":"2024-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142702065","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Effects of magnetic fields on the formation of Interstellar Filaments through shock-cloud interaction","authors":"D. Gogoi,&nbsp;S.M. Borah,&nbsp;E. Saikia","doi":"10.1016/j.ascom.2024.100887","DOIUrl":"10.1016/j.ascom.2024.100887","url":null,"abstract":"<div><div>Interstellar Filaments are ubiquitous in molecular clouds which are hotbeds for star birth. What leads to their formation has been a subject of study in recent years. In the present numerical experiment, we have looked into the role of magnetic field in formation of such structures in the context of multiple molecular cloud complexes after they were subjected to a passing shock. We found that in the absence of this field, post-shock region is turbulent, leading to higher material mixing, 17.5% in the case of the highest porous model considered which also had 42% higher area filling factor compared to models with magnetic field imposed. On the other hand in the presence of a magnetic field, processes such as ‘mass-loading’, slowing down of shock, and inhibition of instabilities are observed which we have found to facilitate the formation of less porous and hence more clumpy structures in post-shock regions. It is found that in the absence of a field, such structures are diffused and spread over a larger area. Such structures are later elongated by hydrodynamical ablation leading to filament-like structures. Morphological output images having filamentary structures are further studied using tools from Nonlinear Dynamics such as Percolation and Fractal Analysis. We find that the filaments formed without a field have higher fractal dimensions, are longer, more complex, and highly branched. Magnetic field influences the properties of the filaments, making them smaller, more confined, and less complex. Further, it is observed that the influence of <strong>B</strong> is diminished with the presence of radiative cooling, still having a subtle affect on the system’s evolution though.</div></div>","PeriodicalId":48757,"journal":{"name":"Astronomy and Computing","volume":"49 ","pages":"Article 100887"},"PeriodicalIF":1.9,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142571724","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Radio frequency interference identification using dual cross-attention and multi-scale feature fusing","authors":"Y. Dao ,&nbsp;B. Liang ,&nbsp;L. Hao ,&nbsp;S. Feng ,&nbsp;S. Wei ,&nbsp;W. Dai ,&nbsp;F. Gu","doi":"10.1016/j.ascom.2024.100881","DOIUrl":"10.1016/j.ascom.2024.100881","url":null,"abstract":"<div><div>Radio astronomy plays a very important role in promoting scientific progress and unraveling the mysteries of the universe. However, radio telescopes are inevitably affected by radio frequency interference (RFI) when receiving radio signals, which leads to a reduction in data quality and has a serious impact on the formation of correct scientific conclusions. Therefore, it is essential to identify the RFI present in the observational data. In order to effectively identify RFI, improve the existing RFI identification methods that suffer from missed detections, and enhance the performance of RFI identification, this paper proposes a novel method that combines a dual cross-attention mechanism with multi-scale feature fusion. Experimental studies were conducted using the observational data from the 40-meter radio telescope at the Yunnan Astronomical Observatory of the Chinese Academy of Sciences. The proposed method achieved scores of 92.49%, 83.90%, and 87.99% in terms of <span><math><mrow><mi>p</mi><mi>r</mi><mi>e</mi><mi>c</mi><mi>i</mi><mi>s</mi><mi>i</mi><mi>o</mi><mi>n</mi></mrow></math></span>, <span><math><mrow><mi>r</mi><mi>e</mi><mi>c</mi><mi>a</mi><mi>l</mi><mi>l</mi></mrow></math></span>, and <span><math><msub><mrow><mi>F</mi></mrow><mrow><mn>1</mn><mo>−</mo><mi>s</mi><mi>c</mi><mi>o</mi><mi>r</mi><mi>e</mi></mrow></msub></math></span>, respectively. It outperformed existing methods (U-Net, RFI-Net, R-Net6, RFI-GAN, EMSCA-UNet) in <span><math><mrow><mi>r</mi><mi>e</mi><mi>c</mi><mi>a</mi><mi>l</mi><mi>l</mi></mrow></math></span> and <span><math><msub><mrow><mi>F</mi></mrow><mrow><mn>1</mn><mo>−</mo><mi>s</mi><mi>c</mi><mi>o</mi><mi>r</mi><mi>e</mi></mrow></msub></math></span>, effectively reducing the occurrence of missed detections and improving the overall performance of radio frequency interference identification.</div></div>","PeriodicalId":48757,"journal":{"name":"Astronomy and Computing","volume":"49 ","pages":"Article 100881"},"PeriodicalIF":1.9,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142528479","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Constraining Galaxy-Halo connection using machine learning","authors":"A. Jana ,&nbsp;L. Samushia","doi":"10.1016/j.ascom.2024.100883","DOIUrl":"10.1016/j.ascom.2024.100883","url":null,"abstract":"<div><div>We investigate the potential of machine learning (ML) methods to model small-scale galaxy clustering for constraining Halo Occupation Distribution (HOD) parameters. Our analysis reveals that while many ML algorithms report good statistical fits, they often yield likelihood contours that are significantly biased in both mean values and variances relative to the true model parameters. This highlights the importance of careful data processing and algorithm selection in ML applications for galaxy clustering, as even seemingly robust methods can lead to biased results if not applied correctly. ML tools offer a promising approach to exploring the HOD parameter space with significantly reduced computational costs compared to traditional brute-force methods if their robustness is established. Using our ANN-based pipeline, we successfully recreate some standard results from recent literature. Properly restricting the HOD parameter space, transforming the training data, and carefully selecting ML algorithms are essential for achieving unbiased and robust predictions. Among the methods tested, artificial neural networks (ANNs) outperform random forests (RF) and ridge regression in predicting clustering statistics, when the HOD prior space is appropriately restricted. We demonstrate these findings using the projected two-point correlation function (<span><math><mrow><msub><mrow><mi>w</mi></mrow><mrow><mi>p</mi></mrow></msub><mrow><mo>(</mo><msub><mrow><mi>r</mi></mrow><mrow><mi>p</mi></mrow></msub><mo>)</mo></mrow></mrow></math></span>), angular multipoles of the correlation function (<span><math><mrow><msub><mrow><mi>ξ</mi></mrow><mrow><mi>ℓ</mi></mrow></msub><mrow><mo>(</mo><mi>r</mi><mo>)</mo></mrow></mrow></math></span>), and the void probability function (VPF) of Luminous Red Galaxies from Dark Energy Spectroscopic Instrument mocks. Our results show that while combining <span><math><mrow><msub><mrow><mi>w</mi></mrow><mrow><mi>p</mi></mrow></msub><mrow><mo>(</mo><msub><mrow><mi>r</mi></mrow><mrow><mi>p</mi></mrow></msub><mo>)</mo></mrow></mrow></math></span> and VPF improves parameter constraints, adding the multipoles <span><math><msub><mrow><mi>ξ</mi></mrow><mrow><mn>0</mn></mrow></msub></math></span>, <span><math><msub><mrow><mi>ξ</mi></mrow><mrow><mn>2</mn></mrow></msub></math></span>, and <span><math><msub><mrow><mi>ξ</mi></mrow><mrow><mn>4</mn></mrow></msub></math></span> to <span><math><mrow><msub><mrow><mi>w</mi></mrow><mrow><mi>p</mi></mrow></msub><mrow><mo>(</mo><msub><mrow><mi>r</mi></mrow><mrow><mi>p</mi></mrow></msub><mo>)</mo></mrow></mrow></math></span> does not significantly improve the constraints.</div></div>","PeriodicalId":48757,"journal":{"name":"Astronomy and Computing","volume":"49 ","pages":"Article 100883"},"PeriodicalIF":1.9,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142698318","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Classifying the clouds of Venus using unsupervised machine learning","authors":"J. Mittendorf ,&nbsp;K. Molaverdikhani ,&nbsp;B. Ercolano ,&nbsp;A. Giovagnoli ,&nbsp;T. Grassi","doi":"10.1016/j.ascom.2024.100884","DOIUrl":"10.1016/j.ascom.2024.100884","url":null,"abstract":"<div><div>Because Venus is completely shrouded by clouds, they play an important role in the planet’s atmospheric dynamics. Studying the various morphological features observed on satellite imagery of the Venusian clouds is crucial to understanding not only the dynamic atmospheric processes, but also interactions between the planet’s surface structures and atmosphere. While attempts at manually categorizing and classifying these features have been made many times throughout Venus’ observational history, they have been limited in scope and prone to subjective bias. We therefore present and investigate an automated, objective, and scalable approach for their classification using unsupervised machine learning that can leverage full datasets of past, ongoing, and future missions.</div><div>To achieve this, we introduce a novel framework to generate nadir observation patches of Venus’ clouds at fixed consistent scales from satellite imagery data of the <em>Venus Express</em> and <em>Akatsuki</em> missions. Such patches are then divided into classes using an unsupervised machine learning approach that consists of encoding the patch images into feature vectors via a convolutional neural network trained on the patch datasets and subsequently clustering the obtained embeddings using hierarchical agglomerative clustering.</div><div>We find that our approach demonstrates considerable accuracy when tested against a curated benchmark dataset of Earth cloud categories, is able to identify meaningful classes for global-scale (3000<!--> <!-->km) cloud features on Venus and can detect small-scale (25<!--> <!-->km) wave patterns. However, at medium scales (<span><math><mo>∼</mo></math></span>500<!--> <!-->km) challenges are encountered, as available resolution and distinctive features start to diminish and blended features complicate the separation of well defined clusters.</div></div>","PeriodicalId":48757,"journal":{"name":"Astronomy and Computing","volume":"49 ","pages":"Article 100884"},"PeriodicalIF":1.9,"publicationDate":"2024-10-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142438005","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}