Yu Sha

📰 News

🔍 Researches

• AI in Industry

  • Supervised learning for acoustic damage detection and diagnosis [1, 2, 3];
  • Unsupervised learning for acoustic signal anomalous sound detection [4].

• Knowledge Guided AI

  • Physics knowledge-guided deep learning [5, 6].

💼 Experiences

📖 Educations

📚 Publications

Papers:

KDD 2024 (Oral&Poster)

Hierarchical Knowledge Guided Fault Intensity Diagnosis of Complex Industrial Systems
Yu Sha, Shuiping Gou, Bo Liu, Ningtao Liu, Johannes Faber, Stefan Schramm, Horst Stoecker, Thomas Steckenreiter, Domagoj Vnucec, Nadine Wetzstein, Andreas Widl, Kai Zhou
SIGKDD 2024 | CCF: A | Paper | Code
• The paper explores hierarchical Knowledge of labels with GCN.
• The paper presents the Hierarchical Knowledge Correlation Matrix.
• The sound cavitation datasets provided by SAMSON AG.
ESWA 2024

Hierarchical Cavitation Intensity Recognition Using Sub-Master Transition Network-based Acoustic Signals in Pipeline Systems
Shuiping Gou, Yu Sha*, Bo Liu, Ningtao Liu, Johannes Faber, Stefan Schramm, Horst Stoecker, Thomas Steckenreiter, Domagoj Vnucec, Nadine Wetzstein, Andreas Widl, Kai Zhou (If you know, you know)
Expert Systems with Applications | 2024 |SCI: 1-Top | Paper
• The paper presents a two-stage hierarchical classification network
• The paper proposes a hierarchical label tree.
• The sound cavitation datasets provided by SAMSON AG.
KDD 2022 (Poster)



Regional-Local Adversarially Learned One-Class Classifier Anomalous Sound Detection in Global Long-Term Space
Yu Sha, Shuiping Gou, Johannes Faber, Bo Liu, Wei Li, Stefan Schramm, Horst Stoecker, Thomas Steckenreiter, Domagoj Vnucec, Nadine Wetzstein, Andreas Widl, Kai Zhou
SIGKDD 2022 | CCF: A | Paper | Code

• The paper proposes a global filter layer based on a 1D FFT global filter for long-term interactions.
• The paper extends the capability of discriminating real against fake signals to differentiating between local and regional reconstructions.
• The paper designs a novel balanceable detection strategy.

EAAI 2022



A multi-task learning for cavitation detection and cavitation intensity recognition of valve acoustic signals
Yu Sha, Johannes Faber, Shuiping Gou, Bo Liu, Wei Li, Stefan Schramm, Horst Stoecker, Thomas Steckenreiter, Domagoj Vnucec, Nadine Wetzstein, Andreas Widl, Kai Zhou
Engineering Applications of Artificial Intelligence | 2022 | SCI: 2-Top | Paper | Code

• The paper regards cavitation detection and cavitation intensity recognition as a multi-task learning.
• The 1-D Double Hierarchical Residual Blocks with large kernel are proposed as an automatic feature extractor.
• The sound cavitation datasets provided by SAMSON AG.

Measurement 2022



An acoustic signal cavitation detection framework based on XGBoost with adaptive selection feature engineering
Yu Sha, Johannes Faber, Shuiping Gou, Bo Liu, Wei Li, Stefan Schramm, Horst Stoecker, Thomas Steckenreiter, Domagoj Vnucec, Nadine Wetzstein, Andreas Widl, Kai Zhou
Measurement | 2022 | SCI: 2-Top | Paper | Code

• The paper proposes a adaptive feature aggregation module.
• The paper presents a adaptive feature crosses module.
• The sound cavitation datasets provided by SAMSON AG.

Front. Neurosci. 2025



Dynamic Spatio-Temporal Pruning for Efficient Spiking Neural Networks
Shuiping Gou, Jiahui Fu, Yu Sha, Zhen Cao, Zhang Guo, Jason K. Eshraghian, Ruimin Li, Licheng Jiao
Frontiers in Neuroscience | 2025 |SCI: 2-Top | Paper

• A spatio-temporal pruning algorithm is proposed to reduce the temporal redundancy,
• Spatial pruning leverages global and inter-layer parameter statistics to minimize model degradation under extreme sparsity.

JBHI 2024



Self-supervised, Non-Contact Heartbeat Detection Based on Ballistocardiograms utilizing Physiological Information Guidance
Changezhe Jiao, Aoyu Yang, Hantao Zhao, Ruhan Yi, Shuiping Gou, Yu Sha, Wanshun Wen, Licheng Jiao, Marjorie Skubic
IEEE Journal of Biomedical and Health Informatics | 2024 |SCI: 1-Top | Paper

• A non-contact, self-supervised heart rate detection method based on physiological information,
• The paper proposes a heartbeat mapping algorithm based on Bidirectional Long Short-Term Memory Network.

MIR 2023



Prioritization Hindsight Experience based on Spatial Position Attention for Robots
Ye Yuan, Yu Sha, Feixiang Sun, Haofan Lu, Shuiping Gou, Jie Luo
Machine Intelligence Research | 2025 | SCI: 4 | Paper

• The paper proposes a spatial position attention module for the existing HER framework.
• The paper presents a theoretical analysis of the total spatial position distance of manipulated object.
• Eight robotic manipulation tasks in the Fetch and Hand robot environments of OpenAI Gym.

DH3D 2022



Phase Retrieval for Terahertz Holography with Physics-Informed Deep Learning
Mingjun Xiang, Lingxiao Wang, Yu Sha, Hui Yuan, Kai Zhou, Hartmut G Roskos
Digital Holography and Three-Dimensional Imaging | 2022 | EI | Paper

• The paper proposes a two novel phase retrieval methods for THz holography.
• The paper employs unsupervised learning and supervised learning based on the MNIST dataset.

AIIG 2022



A study on small magnitude seismic phase identification using 1D deep residual neural network
Wei Li, Megha Chakraborty, Yu Sha, Kai Zhou, Johannes Faber, Georg Rümpker, Horst Stöcker, Nishtha Srivastava
Artificial Intelligence in Geosciences | 2022 | SCI: 2 | Paper | Code

• 1D deep Residual Neural Network for tackling the problem of seismic signal detection and phase identification.
• The method is trained and tested on the dataset recorded by the Southern California Seismic Network.
• The model generalizability is also tested further on the STanford EArthquake Dataset.

Deep Learning-based Small Magnitude Earthquake Detection and Seismic Phase Classification
Wei Li, Yu Sha, Kai Zhou, Johannes Faber, Georg Rümpker, Horst Stöcker, Nishtha Srivastava
arXiv 2204.02870 | Paper | Code

Smart home system based on STC89C52
Liang Qin, Yu Sha, Yumeng Xu
Practical Electronics | 2018 | Paper

Study on operation analysis and decision making for sharing-bicycles
Hong Zhang, Dixin Zhou, Chuanqi Cheng, Yu Sha
Big Data Research | 2019 | Paper

Patents:

1. Heart rate estimation method based on cross-modal mapping of heart impact map signals
   Shuiping Gou, Yuanjie Liu, Ningtao Liu, Yu Sha, Changzhe Jiao, Dong Hai, Shasha Mao, Jiaxin Cheng
   Invention Patent | CN111887858B |

2. Hierarchical Cavitation Intensity Recognition Using Sub-Master Transition Network-based Acoustic Signals in Pipeline Systems
   Shuiping Gou, Yu Sha, Zhang Guo, Bo Liu
   Invention Patent | 202310945885.1 |

3. Face Aging Method Based on Adversarial Learning of Local and Global Region Policies
   Shuiping Gou, Ruichen Xue, Nuo Tong, Ruimin Li, Yu Sha, Ningtao Liu
   Invention Patent | 202310941541.3 |

4. Physics-Guided Global Filtering Attention Convolutional Neural Networks for Image Classification
   Zhang Guo, Huiyu Chen, Shuiping Gou, Yu Sha, Xinlin Wang, Bo Liu
   Invention Patent | 202411474637.4 |

🧰 Projects

Xidian University (XDU):

1. Analysis of battlefield situation of artificial intelligence in the context of big data
   The 20th Research Institute of China Electronics Technology Group Corporation-Xidian Joint Laboratory for Artificial Intelligence | Completed | Project Participant
   

2. Knee intensity recognition for smart wearable devices based on machine learning
   Xijing Hospital | Completed | Project Participant
   

Frankfurt Institute for Advanced Studies (FIAS):

1. Cavitation and leakage detection in large pump/pipe using AI method
   SAMSON AG | Under Study | Project Participant
   

Lanzhou University of Technology (LUT):

1. LoRa based non-contact life detection system
   National Undergraduate Innovation and Entrepreneurship Training Project | Completed | Project Participant
   

2. Evaluation of students' comprehensive ability and program design based on fuzzy theory
   LUT Innovation and Entrepreneurship Training Program | Completed | Project Leader
   

3. Research on the cultivation of college students' creative ability by mathematical modeling competition
   LUT Technology Innovation Fund | Completed | Project Participant
   

4. WeChat program design for second-hand transaction and donation platform on campus
   LUT Technology Innovation Fund | Completed | Project Participant