WISE: A Lightweight, Weakly-Supervised Model for Onboard Fire Smoke Detection and Localization
为解决卫星图像中火烟检测与定位问题，提出WISE模型，采用弱监督蒸馏策略，在资源受限条件下实现高效检测和定位。
0 citationsRead paper
Selected work
Representative Papers
Adaptive Band Selection for Hyperspectral Classification with Spatially Disjoint Evaluation
This work addresses the limitations of existing hyperspectral band selection methods, which are sensitive to initialization, require a preset number of bands lacking flexibility, and exhibit unstable performance under spatially disjoint evaluation protocols. To overcome these issues, the authors propose SGBR-HC, a two-stage approach that first performs supervised band ranking based on class separability and spectral diversity to provide an informative prior for learnable sparse gating. In the second stage, the sparse gating module and a spatial classifier are jointly trained to adaptively determine the optimal number of bands. By integrating differentiable sparse gating, Hard-Concrete initialization, and spatially disjoint evaluation, the method effectively prevents information leakage. Experiments on the Pavia University and Houston 2013 datasets demonstrate state-of-the-art overall accuracy and Cohen’s kappa using only approximately 20 selected bands, while ablation studies confirm the critical role of the ranking prior.
0 citationsRead paper
LLM-Enhanced Semantic Data Integration of Electronic Component Qualifications in the Aerospace Domain
This study addresses the challenge of fragmented qualification data for electronic components in aerospace engineering, which is scattered across multiple heterogeneous systems and impedes efficient decision-making during design phases. To overcome this, the authors propose a semantic integration approach that synergistically combines virtual knowledge graphs with large language models. By leveraging an ontology-based data access (OBDA) framework alongside vector retrieval mechanisms, the method enables unified and efficient querying of qualification information across disparate data silos. The approach maintains strong semantic consistency while substantially reducing manual data curation costs. Compared to conventional retrieval-augmented generation (RAG) or pure large-model solutions, it demonstrates superior performance in retrieval accuracy, computational efficiency, and long-term operational cost, thereby effectively minimizing redundant certification efforts.
0 citationsRead paper
First On-Orbit Demonstration of a Geospatial Foundation Model
This work addresses the challenge of deploying large geospatial foundation models (GeoFMs) on resource-constrained spaceborne hardware. We propose a compact vision transformer architecture integrating model compression, cross-domain adaptation, and multi-task collaborative optimization to significantly reduce computational and memory overhead while preserving generalization capability and accuracy across downstream Earth observation tasks. Our approach enables reliable on-orbit inference for the first time on the IMAGIN-e payload aboard the International Space Station, supporting five core tasks—including land-cover classification and cloud detection. Compared to baseline models, it achieves a 72% reduction in parameter count, a 65% decrease in inference latency, and less than 1.2% accuracy degradation. This work establishes a scalable, flight-verified paradigm for intelligent onboard processing in Earth observation missions.
0 citationsRead paper
Confidence-gated training for efficient early-exit neural networks
Early-exit neural networks suffer from gradient interference during joint training, where deeper classifiers dominate shallow exits and undermine their effectiveness. To address this, we propose Confidence-Gated Training (CGT), a novel training paradigm that introduces conditional gradient propagation: gradients are backpropagated to deeper layers only when the current layer’s prediction confidence falls below a learnable threshold—thereby aligning training dynamics with inference-time early-exit behavior. CGT explicitly grants shallow classifiers priority in decision-making, reducing redundant computation. Extensive experiments on benchmarks including Indian Pines and Fashion-MNIST demonstrate that CGT significantly reduces average inference latency (up to 42%) while improving overall accuracy (+0.8%–1.3%) and early-exit accuracy (+3.1%–5.7%). These gains enhance deployment efficiency in resource-constrained environments without architectural modifications.
0 citationsRead paper
Recent publications
Latest Papers
WISE: A Lightweight, Weakly-Supervised Model for Onboard Fire Smoke Detection and Localization
为解决卫星图像中火烟检测与定位问题，提出WISE模型，采用弱监督蒸馏策略，在资源受限条件下实现高效检测和定位。
0 citationsRead paper
Adaptive Band Selection for Hyperspectral Classification with Spatially Disjoint Evaluation
This work addresses the limitations of existing hyperspectral band selection methods, which are sensitive to initialization, require a preset number of bands lacking flexibility, and exhibit unstable performance under spatially disjoint evaluation protocols. To overcome these issues, the authors propose SGBR-HC, a two-stage approach that first performs supervised band ranking based on class separability and spectral diversity to provide an informative prior for learnable sparse gating. In the second stage, the sparse gating module and a spatial classifier are jointly trained to adaptively determine the optimal number of bands. By integrating differentiable sparse gating, Hard-Concrete initialization, and spatially disjoint evaluation, the method effectively prevents information leakage. Experiments on the Pavia University and Houston 2013 datasets demonstrate state-of-the-art overall accuracy and Cohen’s kappa using only approximately 20 selected bands, while ablation studies confirm the critical role of the ranking prior.
0 citationsRead paper
LLM-Enhanced Semantic Data Integration of Electronic Component Qualifications in the Aerospace Domain
This study addresses the challenge of fragmented qualification data for electronic components in aerospace engineering, which is scattered across multiple heterogeneous systems and impedes efficient decision-making during design phases. To overcome this, the authors propose a semantic integration approach that synergistically combines virtual knowledge graphs with large language models. By leveraging an ontology-based data access (OBDA) framework alongside vector retrieval mechanisms, the method enables unified and efficient querying of qualification information across disparate data silos. The approach maintains strong semantic consistency while substantially reducing manual data curation costs. Compared to conventional retrieval-augmented generation (RAG) or pure large-model solutions, it demonstrates superior performance in retrieval accuracy, computational efficiency, and long-term operational cost, thereby effectively minimizing redundant certification efforts.
0 citationsRead paper
First On-Orbit Demonstration of a Geospatial Foundation Model
This work addresses the challenge of deploying large geospatial foundation models (GeoFMs) on resource-constrained spaceborne hardware. We propose a compact vision transformer architecture integrating model compression, cross-domain adaptation, and multi-task collaborative optimization to significantly reduce computational and memory overhead while preserving generalization capability and accuracy across downstream Earth observation tasks. Our approach enables reliable on-orbit inference for the first time on the IMAGIN-e payload aboard the International Space Station, supporting five core tasks—including land-cover classification and cloud detection. Compared to baseline models, it achieves a 72% reduction in parameter count, a 65% decrease in inference latency, and less than 1.2% accuracy degradation. This work establishes a scalable, flight-verified paradigm for intelligent onboard processing in Earth observation missions.
0 citationsRead paper
Confidence-gated training for efficient early-exit neural networks
Early-exit neural networks suffer from gradient interference during joint training, where deeper classifiers dominate shallow exits and undermine their effectiveness. To address this, we propose Confidence-Gated Training (CGT), a novel training paradigm that introduces conditional gradient propagation: gradients are backpropagated to deeper layers only when the current layer’s prediction confidence falls below a learnable threshold—thereby aligning training dynamics with inference-time early-exit behavior. CGT explicitly grants shallow classifiers priority in decision-making, reducing redundant computation. Extensive experiments on benchmarks including Indian Pines and Fashion-MNIST demonstrate that CGT significantly reduces average inference latency (up to 42%) while improving overall accuracy (+0.8%–1.3%) and early-exit accuracy (+3.1%–5.7%). These gains enhance deployment efficiency in resource-constrained environments without architectural modifications.
0 citationsRead paper