Chemically peculiar (CP) stars offer a unique observational sample for exploring key processes in stellar evolution. Leveraging the massive amounts of data obtained from large-scale sky surveys, deep learning technology has significantly advanced research in identifying special celestial objects in large datasets in recent years. In this study, we developed a hybrid deep learning approach that integrates a long short-term memory network with a convolutional neural network to search for CP1/CP2/CP3 stars in low-resolution spectra of LAMOST DR12 (v1.0). We first trained a binary classification model to distinguish CP stars from non-CP stars, achieving an accuracy of 95.92%. Applying this model to the LAMOST DR12 (v1.0) dataset, we identified 79,216 CP star candidates. We then constructed a three-class classification model to distinguish between CP1, CP2, and CP3, achieving an accuracy rate of 97.34%. This model further identified 59,333 CP1 stars, 15,179 CP2 stars, and 1424 CP3 stars from the CP star candidates. After cross validation, 51,241 stars were consistent with previously reported CP stars based on crossmatching results, with 24,695 newly identified CP1/CP2/CP3 candidates, including 16,432 new CP1 candidates, 7132 CP2 candidates, and 1131 CP3 candidates. By combining visual inspection with the MKCLASS code, we ultimately identified 2899 CP1 stars, 4536 CP2 stars, and 747 CP3 stars.

Be stars are rapidly rotating B-type stars that exhibit Balmer emission lines in their optical spectra, which makes them crucial for studying stellar evolution and circumstellar disk structures. In this study, we performed a systematic identification of Be stars using low-resolution spectroscopic data from Large Sky Area Multi-Object Fiber Spectroscopic Telescope Data Release 11. We constructed a dataset and developed a hybrid identification model combining long short-term memory networks and convolutional neural networks, achieving a testing accuracy of 97.83%. The trained model was applied to spectra with signal-to-noise ratios greater than 10 in the g band, yielding 55,667 B-type candidates. After further validation using the MKCLASS tool, manual verification, and removing duplicated observational data, 26,925 B-type candidates were confirmed. These B-type candidates were subsequently crossmatched with published H alpha emission-line catalogs to identify emission-line B-type objects, yielding a sample of 5384 Be star candidates. By comparing these samples with existing Be star databases, we identified 2881 previously reported objects and 2503 newly discovered Be stars. Based on infrared color criteria, 1658 of these new detections were classified as classical Be stars and 32 as Herbig Be stars. The remaining 813 objects were categorized as inconclusive, as they either exhibited ambiguous classifications satisfying only partial infrared criteria or lacked sufficient photometric information for definitive categorization.

Very metal-poor (VMP; [Fe/H] < -2) stars are critical tracers for understanding early star formation and Galactic chemical evolution. However, identifying these rare objects from the massive datasets generated by the Dark Energy Spectroscopic Instrument (DESI) presents significant challenges in efficiency and precision due to the scarcity of high-fidelity labels and low signal-to-noise ratios in the metal-poor regime. To address this, we propose a novel dual-model deep learning framework that integrates a 1D-ResNet binary classifier with a specialized parameter regression model. Leveraging a transfer learning strategy with high-quality labels from APOGEE and the Large Sky Area Multi-object Fiber Spectroscopic Telescope, we optimized the framework for DESI spectra. The classification model achieves an accuracy of 97.87%, while the regression model predicts stellar metallicity with an rms error of 0.093 dex. Applying this framework to the high-quality DESI-HQ-DATA dataset and applying a strict temperature cut (T-eff >= 4500 K) to avoid extrapolation in the cool dwarf regime, we constructed a highly purified catalog of 2569 high-confidence VMP candidates. Internal validation demonstrates a significant improvement in consistency between DESI pipeline measurements, and external crossmatching with Gaia XP spectra confirms the reliability of our metallicity estimates. Finally, compared to existing compilations, this work contributes 1377 new VMP candidates, providing a robust and statistically significant sample for future studies of the Galactic halo and ancient stellar populations.

White dwarfs (WDs) with infrared (IR) excesses probe dusty debris disks, low-mass companions, and the late-stage evolution of planetary and binary systems. Conventional searches usually rely on source-by-source spectral energy distribution (SED) fitting and visual inspection, which become time-consuming for the rapidly growing samples produced by large spectroscopic surveys. We develop a supervised multimodal deep learning framework for scalable preselection of IR-excess WD candidates in the Dark Energy Spectroscopic Instrument (DESI) Data Release 1. The model combines Pan-STARRS1 z - and y -band images, unWISE W1- and W2-band images, and atmospheric, astrometric, and photometric tabular features. On the internal validation set, the model achieved an area under the receiver operating characteristic curve of 0.9765, demonstrating effective separation of literature-reported IR-excess candidates from comparison WDs. Applied to the 10,988 objects in DESI-WD-SEARCH-DATA, the model selected 1886 first-stage candidates for subsequent validation. Image-based screening for Wide-field Infrared Survey Explorer (WISE)-scale contamination and composition-dependent SED validation identified 1041 objects satisfying the adopted excess criteria. Of these, 294 had sufficient photometric coverage for further assessment, and catalog-specific photometric-quality screening yielded a final catalog of 221 candidates, including 204 main-sample and 17 warning candidates. We also provide the complete list of 1886 first-stage candidates with flags recording the outcomes of subsequent screening steps. This catalog provides targets for future high-resolution IR imaging, spectroscopic follow-up, and studies of the physical origins of IR excesses around WDs.

The H α emission line commonly appears in the spectra of many stars and serves as a key indicator for tracing ionized interstellar gas, investigating stellar activity, and studying gas dynamics. Young stellar objects (YSOs), representing the early evolutionary stages of stars, typically exhibit the H α emission line in their spectra. In this paper, we use bidirectional long short-term memory networks and convolutional neural networks to identify H α emission-line stars in medium-resolution spectra from the Large Area Multi-Target Fiber Optic Spectroscopic Telescope (LAMOST) survey, and further search for YSO candidates via the Li absorption line. We constructed a data set by crossmatching previously published data sets with LAMOST data and performing manual verification. Using this data set, we built an identification model that achieved an accuracy of 97.58% on the testing set. Application of this model to the full survey yielded 46,867 H α emission-line star candidates, with 41,996 visually confirmed detections (15,329 of which are recorded in SIMBAD). To further identify YSOs, we developed a dedicated Li absorption line detector, identifying 4618 preliminary candidates from the H α emission-line stars. Rigorous vetting confirmed 4255 YSO candidates, comprising 3470 previously cataloged objects and 785 new discoveries. All catalogs (H α emission-line stars and YSOs) and the code of the proposed model are publicly released to facilitate community research.

Be stars are rapidly rotating B-type stars that exhibit Balmer emission lines in their optical spectra. These stars play an important role in studies of stellar evolution and disk structures. In this work, we carried out a systematic search for Be stars based on LAMOST spectroscopic data. Using low-resolution spectra from LAMOST DR11, we constructed a data set and developed a classification model that combines long short-term memory networks and convolutional neural networks , achieving a testing accuracy of 97.86%. The trained model was then applied to spectra with signal-to-noise ratios greater than 10, yielding 55,667 B-type candidates. With the aid of the MKCLASS automated classification tool and manual verification, we finally confirmed 40,223 B-type spectra. By cross-matching with published Hα emission-line star catalogs, we obtained a sample of 8298 Be stars, including 3787 previously reported Be stars and 4511 newly discovered. Furthermore, by incorporating color information, we classified the Be star sample into Herbig Be stars and Classical Be stars. In total, we identified 3363 Classical Be stars and 35 Herbig Be stars. The B-type and Be star catalogs derived in this study, together with the code used for model training, have been publicly released to facilitate community research.

The performance of the deconvolution algorithm plays a crucial role in data processing of radio interferometers.The multi-scale multi-frequency synthesis(MSMFS)CLEAN is a widely used deconvolution algorithm for radio interferometric imaging,which combines the advantages of both wide-band synthesis imaging and multi-scale imaging and can substantially improve performance.However,how best to effectively determine the optimal scale is an important problem when implementing the MSMFS CLEAN algorithm.In this study,we proposed a Gaussian fitting method for multiple sources based on the gradient descent algorithm,with consideration of the influence of the point spread function(PSF).After fitting,we analyzed the fitting components using statistical analysis to derive reasonable scale information through the model parameters.A series of simulation validations demonstrated that the scales extracted by our proposed algorithm are accurate and reasonable.The proposed method can be applied to the deconvolution algorithm and provide modeling analysis for Gaussian sources,offering data support for source extraction algorithms.

The Chinese Space Station Telescope (abbreviated as CSST) is a future advanced space telescope. Real-time identification of galaxy and nebula/star cluster (abbreviated as NSC) images is of great value during CSST survey. While recent research on celestial object recognition has progressed, the rapid and efficient identification of high-resolution local celestial images remains challenging. In this study, we conducted galaxy and NSC image classification research using deep learning methods based on data from the Hubble Space Telescope. We built a Local Celestial Image Dataset and designed a deep learning model named HR-CelestialNet for classifying images of the galaxy and NSC. HR-CelestialNet achieved an accuracy of 89.09% on the testing set, outperforming models such as AlexNet, VGGNet and ResNet, while demonstrating faster recognition speeds. Furthermore, we investigated the factors influencing CSST image quality and evaluated the generalization ability of HR-CelestialNet on the blurry image dataset, demonstrating its robustness to low image quality. The proposed method can enable real-time identification of celestial images during CSST survey mission.

Identifying and classifying variable stars is essential to time-domain astronomy. The Large Area Multi-Object Fiber Optic Spectroscopic Telescope (LAMOST) acquired a large amount of spectral data. However, there is no corresponding variable source-related information in the data, constraining LAMOST data utilization for scientific research. In this study, we systematically investigated variable source classification methods for LAMOST data. We constructed a 10-class classification model using three mainstream machine-learning methods. Through performance comparison, we chose the LightGBM and XGBoost models. We further identified variable source candidates in the r band in LAMOST DR9 and obtained 281,514 variable source candidates with probabilities greater than 95%. Subsequently, we filtered out the sources of periodic variable sources using the generalized Lomb–Scargle periodogram and classified these periodic variable sources using the classification model. Finally, we propose a reliable periodic variable star catalog containing 176,337 stars with specific types.

Young stellar objects (YSOs) represent the earliest stage in the process of star formation, offering insights that contribute to the development of models elucidating star formation and evolution. Recent advancements in deep-learning techniques have enabled significant strides in identifying special objects within vast data sets. In this paper, we present a YSO identification method based on deep-learning principles and spectra from the LAMOST. We designed a structure based on a long short-term memory network and a convolutional neural network and trained different models in two steps to identify YSO candidates. Initially, we trained a model to detect stellar spectra featuring the H α emission line, achieving an accuracy of 98.67%. Leveraging this model, we classified 10,495,781 stellar spectra from LAMOST, yielding 76,867 candidates displaying a H α emission line. Subsequently, we developed a YSO identification model, which achieved a recall rate of 95.81% for YSOs. Utilizing this model, we further identified 35,021 YSO candidates from the H α emission-line candidates. Following cross validation, 3204 samples were identified as previously reported YSO candidates. We eliminated samples with low signal-to-noise ratios and M dwarfs by using the equivalent widths of the N ii and He i emission lines and visual inspection, resulting in a catalog of 20,530 YSO candidates. To facilitate future research endeavors, we provide the obtained catalogs of H α emission-line star candidates and YSO candidates along with the code used for training the model.

Carbon stars play a crucial role in astronomical research and are significant for understanding stellar evolution, measuring cosmic distances, and studying galaxy kinematics. In recent years, identifying carbon stars using machine learning methods and traditional line-index methods has become a research hotspot, but there are still limitations regarding accuracy and automation. In this study, we propose to build a five-class model to identify carbon stars using spectral data from LAMOST DR9. The model achieved 99.45% precision and 91.21% recall on the carbon star testing set. We conducted independent tests using a sample of 1333 known carbon stars that were not used in the training and testing phases, and our model ultimately identified 1199 carbon stars. On this basis, we used this model to screen 11,226,252 spectra of LAMOST DR9 and identified 4383 carbon stars, including 1197 newly discovered carbon stars. To gain a more comprehensive understanding of the characteristics of the 4383 carbon stars obtained, further visual inspection of these spectra was performed to provide more detailed carbon star subtypes.

Double-line spectroscopic binaries (SB2s) are a vital class of spectroscopic binaries for studying star formation and evolution. Searching for SB2s has been a hot topic in astronomy. Although considerable efforts have been made with fruitful outcomes, limitations in automation and accuracy still persist. In this study, we developed a convolutional neural network model to search for SB2 candidates in LAMOST medium-resolution survey (MRS) data release (DR) 9 v1.0 by detecting double peaks in the cross-correlation function (CCF). We first generated a large number of spectra of single stars and binaries using the iSpec spectral synthesis software. The CCFs of these synthesized spectra were then calculated to form our training set. To efficiently detect the peaks of the CCFs, we applied a Softmax function-based noise reduction method. After testing and validation, the model achieved an accuracy of 97.76% in the testing set and was validated for more than 90% of the sample in several published SB2 catalogs. Finally, by applying the model to examine approximately 1.59 million LAMOST-MRS DR9 spectra, we identified 728 candidate SB2s, including 281 newly discovered ones.

White dwarfs represent the ultimate stage of evolution for over 97% of stars and play a crucial role in studies of the Milky Way's structure and evolution. Recent years have witnessed significant progress in using deep-learning methods for identifying unique objects in large-scale data. In this paper, we present a model based on transfer learning for identifying white dwarfs. We constructed a data set using the spectra released by LAMOST DR9 and trained a convolutional neural network model. The model was then further trained using a transfer-learning approach for a binary classification model. Our final model is comprised of a seven-class classification model and a binary classification model. The testing set yielded an accuracy rate of 96.08%. Our proposed model successfully identifies 4314 of the 4479 white dwarfs published in previous papers. We applied this model to filter the 1,121,128 spectral data from the LAMOST DR9 V1 catalog. Subsequently, we obtained 6317 white dwarf candidates, of which 5014 were cross-validated and found to be known white dwarfs. We finally identified 489 new white dwarfs out of the remaining 1303 candidates, containing 377 DAs, 1 DB, 4 DZs, 1 magnetic WD, 101 DA+M binaries, and 1 DB+M binary. Our study also compared transfer-learning methods with non-transfer-learning methods, and the results show that transfer learning provides faster training speed and a higher accuracy rate. We provide the trained model and a corresponding usage program for subsequent studies.

Hot subdwarf stars are a particular type of star that is crucial for studying binary evolution and atmospheric diffusion processes. In recent years, identifying hot subdwarfs by machine-learning methods has become a hot topic, but there are still limitations in automation and accuracy. In this paper, we proposed a robust identification method based on a convolutional neural network. We first constructed the data set using the spectral data of LAMOST DR7-V1. We then constructed a hybrid recognition model including an eight-class classification model and a binary classification model. The model achieved an accuracy of 96.17% on the testing set. To further validate the accuracy of the model, we selected 835 hot subdwarfs that were not involved in the training process from the identified LAMOST catalog (2428, including repeated observations) as the validation set. An accuracy of 96.05% was achieved. On this basis, we used the model to filter and classify all 10,640,255 spectra of LAMOST DR7-V1, and obtained a catalog of 2393 hot subdwarf candidates, of which 2067 have been confirmed. We found 25 new hot subdwarfs among the remaining candidates by manual validation. The overall accuracy of the model is 87.42%. Overall, the model presented in this study can effectively identify specific spectra with robust results and high accuracy, and can be further applied to the classification of large-scale spectra and the search for specific targets.

Improving the performance of solar flare forecasting is a hot topic in the solar physics research field. Deep learning has been considered a promising approach to perform solar flare forecasting in recent years. We first used the generative adversarial networks (GAN) technique augmenting sample data to balance samples with different flare classes. We then proposed a hybrid convolutional neural network (CNN) model (M) for forecasting flare eruption in a solar cycle. Based on this model, we further investigated the effects of the rising and declining phases for flare forecasting. Two CNN models, i.e., M rp and M dp, were presented to forecast solar flare eruptions in the rising phase and declining phase of solar cycle 24, respectively. A series of testing results proved the following. (1) Sample balance is critical for the stability of the CNN model. The augmented data generated by GAN effectively improved the stability of the forecast model. (2) For C-class, M-class, and X-class flare forecasting using Solar Dynamics Observatory line-of-sight magnetograms, the means of the true skill statistics (TSS) scores of M are 0.646, 0.653, and 0.762, which improved by 20.1%, 22.3%, and 38.0% compared with previous studies. (3) It is valuable to separately model the flare forecasts in the rising and declining phases of a solar cycle. Compared with model M, the means of the TSS scores for No-flare, C-class, M-class, and X-class flare forecasting of the M rp improved by 5.9%, 9.4%, 17.9%, and 13.1%, and those of the M dp improved by 1.5%, 2.6%, 11.5%, and 12.2%.