Wenhua Shao

Dr. Wenhua Shao


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
21 May 2025 in Remote Sensing
Verified
Journal Article
Open access
PDF file
BNBokun Ning
Bokun Ning
FZFang Zhao
Fang Zhao
LHLUO HAIYONG
LUO HAIYONG
DLDan Luo
Dan Luo
WSWenhua Shao
Wenhua Shao
The combination of GNSS RTK and INS offers complementary advantages but faces significant challenges in urban canyons. Frequent cycle slips in carrier phase measurements and ambiguity resolution algorithms increase computational burden without improving positioning accuracy. Additionally, environmental interference introduces noise into observations, potentially leading to complete signal loss. To address these issues, this paper proposes a factor graph optimization (FGO) positioning algorithm incorporating predictive observation factors. First, a penalized spline (P-spline) is constructed to predict and smooth Doppler measurements. The predicted Doppler is then fused with the dynamics model predictions to enhance robustness. Using predictive Doppler, carrier phase and pseudorange observations are reconstructed, generating predictive constraint factors to improve positioning accuracy. Real-world tests conducted in urban canyons, including Shanghai, demonstrate that the proposed method maintains stable positioning performance even under short-term signal outages, effectively mitigating cumulative positioning errors caused by data loss. Compared to traditional methods that rely solely on available observations, the proposed algorithm improves northward and dynamic positioning accuracy by 35% and 29%, respectively, providing a highly robust navigation solution for complex urban environments.

12 May 2025 in Transactions on Artificial Intelligence
Verified
Journal Article
Open access
PDF file
YZYutong Zhao
Yutong Zhao
JPJianye Pang
Jianye Pang
XZXinjie Zhu
Xinjie Zhu
WSWenhua Shao
Wenhua Shao
Traditional automated machine learning (AutoML) often faces limitations in manual effort, complexity management, and subjective design choices. This paper introduces a novel LLM-driven AutoML framework centered on the innovation of decomposed prompting. We hypothesize that by strategically breaking down complex AutoML tasks into sequential, guided sub-prompts, Large Language Models (LLMs) operating within a code sandbox on standard PCs can autonomously design, implement, evaluate, and select high-performing machine learning models. To validate this, we primarily applied our decomposed prompting approach to sleep disorder classification (illustrating potential benefits in healthcare). To assess the generalizability and robustness of our method across different data types, we subsequently evaluated it on the established 20 Newsgroups text classification benchmark. We rigorously compared decomposed prompting against zero-shot and few-shot prompting strategies, as well as a manually engineered baseline. Our results demonstrate that decomposed prompting significantly outperforms these alternatives. Our results demonstrate that decomposed prompting significantly outperforms alternatives, enabling the LLM to autonomously achieve superior classifier design and performance, particularly showing strong results in the primary sleep disorder domain and demonstrating robustness in the benchmark task. These findings underscore the transformative potential of decomposed prompting as a key technique for advancing LLM-driven AutoML across diverse application areas beyond the specific examples explored here, paving the way for more automated and accessible problem-solving in scientific and engineering disciplines.