Mingfu Shao

Dr. Mingfu Shao

orcid

Research Focus & Metrics

Academic Career Length

2Years (2025-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
1 April 2026 in Research in Astronomy and Astrophysics
Verified
Journal Article
Closed access
MSMingfu Shao
Mingfu Shao
HWHui Wang
Hui Wang
LTLiyue Tong
Liyue Tong
YLYuyang Li
Yuyang Li
CWCunshi Wang
Cunshi Wang
JHJing Huang
Jing Huang
The Vision-Language Models (VLMs) have achieved breakthrough progress in general knowledge domains, yet adaptation to specialized scientific fields remains challenging due to multimodal representation shifts and the limited integration of domain-specific knowledge. To address the limitations of general-purpose VLMs when applied to solar physics image recognition, analysis, and reasoning, we propose the first vision–language model specifically designed for solar physics, referred to as JinWu Vision–Language (JW-VL). The model integrates multi-wavelength observational data from both space-based and ground-based telescopes, encompassing representative spectral bands spanning the photosphere, chromosphere, and corona. Built upon a cross-modal alignment knowledge distillation framework, JW-VL learns a joint visual–semantic embedding that enables end-to-end modeling from raw solar observational data to downstream tasks, including solar image recognition, solar activity analysis via image-based question answering, and optical character recognition (OCR), while also supporting the construction of a multi-band, cross-instrument solar image benchmark dataset. Furthermore, as a demonstration of interdisciplinary applicability, we developed a “Daily Solar Activity Reports” agent comprising core modules for solar activity level assessment, significant active region characterization, magnetic field complexity analysis, potential space weather impact assessment, and recommended observational focus activate regions. By enabling multimodal reasoning over multi-band solar observational data, JW-VL bridges raw observations and diverse downstream tasks, and provides a reusable methodological framework for vision–language modeling in solar physics.

12 November 2025 in arXiv
Verified
Preprint
Open access
PDF file
MSMingfu Shao
Mingfu Shao
HWHui Wang
Hui Wang
YLYuyang Li
Yuyang Li
JLJia-Ben Lin
Jia-Ben Lin
JLJifeng Liu
Jifeng Liu
YZYuqing Zhou
Yuqing Zhou
Solar flares, the most powerful explosive phenomena in the solar system, may pose significant hazards to spaceborne satellites and ground-based infrastructure. Despite decades of intensive research, reliable flare prediction remains a challenging task. Large Language Models, as a milestone in artificial intelligence, exhibit exceptional general knowledge and next-token prediction capabilities. Here we introduce JW-Flare, the first Multimodal Large Language Models (MLLMs) explicitly trained for solar flare forecasting through fine-tuning on textual physic parameters of solar active regions and magnetic field images. This method demonstrates state-of-the-art (SOTA) performance for large flares prediction on the test dataset. It effectively identifies all 79 X-class flares from 18,949 test samples, yielding a True Skill Statistic (TSS) of 0.95 and a True Positive Rate (TPR) of 1.00, outperforming traditional predictive models. We further investigate the capability origins of JW-Flare through explainability experiments, revealing that solar physics knowledge acquired during pre-training contributes to flare forecasting performance. Additionally, we evaluate models of different parameter scales, confirming the Scaling_Law of Large Language Models in domain-specific applications, such as solar physics. This study marks a substantial advance in both the scale and accuracy of solar flare forecasting and opens a promising avenue for AI-driven methodologies in broader scientific domains.

26 September 2026 in Universe
Verified
Journal Article
Open access
PDF file
HWHui Wang
Hui Wang
JLJia-Ben Lin
Jia-Ben Lin
MSMingfu Shao
Mingfu Shao
LTLi-Yue Tong
Li-Yue Tong
CYChen Yang
Chen Yang
YLYu-Yang Li
Yu-Yang Li
Ground-based whole-sky imagers provide cloud cover data to determine solar observing conditions. Deep segmentation models exceed 0.87 mean intersection-over-union (IoU, shared area over union) on public benchmarks, but need not transfer to an operating site. We release CLOUD99, a dataset of 99 annotated whole-sky images from the Ganyu Solar Observation Station, a humid coastal site, with an evaluation protocol that excludes invalid and saturated pixels. A U-Net reaches 0.933 cloud IoU on the WSISEG benchmark, against 0.873 for a normalized red–blue ratio threshold. At Ganyu, the same network reaches 0.557 and the same threshold 0.632, so the ranking of the two methods inverts. Self-calibrating the fisheye projection localizes the error near the Sun. The region within 30° of the Sun holds 14.6% of the valid field but 34.8% of the threshold’s false alarms and 49.7% of the network’s misses. They fail in opposite directions, and the difference matters operationally: the threshold detects all 70 annotated solar obscurations, whereas the zero-shot network misses 48. Ten to twenty labelled frames from the site bring the network to parity with the threshold. Benchmark accuracy alone is therefore not a sufficient basis for choosing a cloud segmentation method for automated solar observing.

13 July 2026 in Research in Astronomy and Astrophysics
Verified
Journal Article
Closed access
HWHui Wang
Hui Wang
MSMingfu Shao
Mingfu Shao
YLYuyang Li
Yuyang Li
JLJia Ben Lin
Jia Ben Lin
LTLiyue Tong
Liyue Tong
CYChen Yang
Chen Yang
Accurate sunspot classification is essential for assessing the eruptive potential of solar active regions and forecasting space weather. We present JW-SSD, a high-quality multimodal benchmark dataset for fine-grained magnetic-type classification of sunspots. Constructed from SDO/HMI SHARP 720,s data (2010--2023, Solar Cycles~24 and~25), JW-SSD comprises 36,553 co-registered magnetogram--continuum pairs from 2,507 active regions. Unlike conventional three-class schemes, JW-SSD refines the Mount Wilson classification into five physically meaningful categories ($\alpha$, $\beta$, $\beta$-$\delta$, $\beta$-$\gamma$, $\beta$-$\gamma$-$\delta$), enabling finer characterization of magnetic complexity. Rigorous quality control---including central meridian distance restriction, saturation filtering, and sharpness screening---ensures high data validity. The dataset is provided in both FITS and PNG formats, with standard training (29,243) and test (7,310) splits. Benchmark experiments with four representative architectures (U-Net, ResNet-50, EfficientNet-B0, and ViT-Small) yield high accuracy across all models (89.43%--94.78% on the three-class task), confirming that the dataset is reliably learnable across diverse modeling paradigms. JW-SSD has further been employed to train JW-SunSpot, a multimodal large language model that achieves the highest classification accuracy, demonstrating the dataset's broad applicability to both conventional networks and large-language-model-based approaches.

16 September 2026 in Universe
Verified
Journal Article
Open access
PDF file
MSMingfu Shao
Mingfu Shao
JLJia-Ben Lin
Jia-Ben Lin
HWHui Wang
Hui Wang
LTLi-Yue Tong
Li-Yue Tong
CYChen Yang
Chen Yang
YLYu-Yang Li
Yu-Yang Li
Solar flares drive severe space weather hazards, and forecasting their occurrence remains a central challenge for both heliophysics and operational space weather services. Data-driven methods require long-horizon datasets in which images, magnetic features, and flare labels are coregistered in space and time. We present JW-FD (JinWu-Flare Dataset), a 15-year multimodal release spanning 1 January 2011 through 31 December 2025, constructed from SDO/HMI line-of-sight magnetograms, NOAA Solar Region Summary reports, and NOAA X-ray flare event lists. The dataset comprises 3,064 independent active regions and 1,991,247 coregistered magnetogram crops, together with FITS, PNG, CSV, and MP4 modalities. Each sample provides 29 magnetic features linked to configurable flare labels under a strict pre-eruption window spanning seven forecast horizons and four GOES intensity thresholds. An 8:1:1 split at the active region level is adopted to prevent temporal leakage between partitions. PNG branches are released at ten magnetic saturation thresholds, and internal Transformer and four-model 24 h experiments on ≥C1.0 and ≥M1.0 forecasting suggest Bth=1000 G as a preliminary default, although skill remains modest and the optimal saturation is model- and task-dependent. The open-source construction pipeline and dataset are publicly available.