Ying Zhu

EAAI Journal 2026 Journal Article

A hierarchical domain adaptive explainable vulnerability detection scheme for power systems based on dynamic slicing enhancement

• Fangfang Dang
• Jiyu Zhang
• Kehe Wu
• Shuai Li
• Ying Zhu

Automated vulnerability detection using graph neural networks has demonstrated significant potential, yet models trained on general-source software often exhibit high false-positive rates when applied to specialized power system code due to severe domain shift. To bridge this gap, this paper proposes a novel hierarchical domain adaptive and explainable vulnerability detection framework (HDAvul), specifically designed for power systems, whose core innovation lies in three key components. First, a dynamic slicing mechanism enhanced with power-dedicated semantic edges is introduced to adaptively capture the complete functional context of vulnerabilities, thereby overcoming the semantic fragmentation inherent in traditional fixed-window methods. Second, a hierarchical domain adaptation module is developed, employing multi-granularity adversarial learning to align feature distributions simultaneously at the Node, Slice, and Graph levels to ensure the preservation of intricate structural logic during knowledge transfer. Third, a transparent explanation module is integrated to automatically generate structural vulnerability-triggering paths (VTPs), providing traceable evidence for expert validation in safety-critical environments. Experimental results on real-world power system projects demonstrate that HDAvul achieves a peak F1-Score of 85. 2%, outperforming state-of-the-art domain-free and single-level adaptation baselines, while achieving a superior explanation stability of 89. 0% that confirms its capacity to provide reliable, high-fidelity analysis for securing mission-critical power infrastructure.

JBHI Journal 2026 Journal Article

Leads-Adaptive Fetal Electrocardiogram Extraction Using Attention-Based BiLSTM

• Ying Zhu
• Le Xu
• Shenao Chen
• Limin Diao
• Yi Zheng
• Varun G Menon

Extracting the fetal electrocardiogram (fECG) from the abdominal electrocardiogram (AECG) will help clinicians accurately discern fetal cardiac rhythm patterns. Nevertheless, the intricacies inherent in the clinical setting often precipitate signal anomalies within AECG recordings, thereby rendering traditional extraction methodologies suboptimal. This work proposes a deep learning-based fECG extraction method designed for the adaptive extraction of fECG signals from multi-lead AECG inputs. The proposed methodology is predicated upon a bidirectional long short-term memory (BiLSTM) architecture, augmented with a deep supervision subnetwork and an attention mechanism module. The attention module quantifies the inter-channel relevance of the input AECG through the computation of attention weights, facilitating subsequent feature fusion along the channel axis. This process effectively mitigates the impact of defective channels on the output. Comprehensive evaluations were conducted on publicly available datasets, encompassing scenarios with channel loss and benchmarked against established models. The results demonstrate the proposed model's efficacy in extracting fECG signals under diverse channel defect conditions. Ablation studies further validate the critical role of the attention module in enhancing the model's resilience to channel anomalies. Additionally, the reliability of the extracted fECG signals was corroborated through experiments involving input signal masking. The method proposed in this work is helpful for the clinical deployment of the fECG extraction in fetal cardiac rhythm assessment.

EAAI Journal 2025 Journal Article

A digital twin-assisted algorithm for diagnosis of permanent magnet synchronous generator interturn short circuit fault and converter open circuit fault in wind power systems using Pearson correlation coefficient

• Bin Sun
• Ying Zhu
• Zhinong Wei

Interturn short-circuit faults (ISCFs) in permanent magnet synchronous generators (PMSGs) and open-circuit faults (OCFs) in the machine-side converters represent two critical reliability challenges in wind power systems. Conventional fault diagnosis approaches typically rely on dedicated models for each fault type for each fault type, leading to excessive system complexity and suboptimal computational efficiency. To overcome these limitations, this paper proposes a novel unified digital twin-assisted framework capable of simultaneous diagnosis of both PMSG ISCFs and converter OCFs within a single integrated architecture. The high-fidelity digital twin model based on one-dimensional convolutional neural networks is established to generate real-time reference value of current space vector (SV) for online fault detection, while Pearson correlation coefficient analysis enables accurate differentiation between ISCF and OCF. For ISCFs, the fault severity assessment is performed based on the deviation between reference and measured current SV, with the faulty phase identified using phase current root mean square (RMS) values. In the case of converter OCFs, the proposed method introduces a dual-stage identification process: single and dual insulated-gate bipolar transistor (IGBT) open faults are differentiated through severity estimation analysis, and the faulty IGBT is identified by evaluating the effective current interval ratio (ECIR) and normalized current average (NCA). The experimental results validate the effectiveness of the proposed method, and comparative analysis further demonstrates its superior performance in terms of parameter dependency and diagnostic efficacy.

IROS Conference 2025 Conference Paper

TCNet: A Temporally Consistent Network for Self-supervised Monocular Depth Estimation

• Ying Zhu
• Hong Liu
• Jianbing Wu
• Mengyuan Liu

Despite significant advances in self-supervised monocular depth estimation methods, achieving temporally consistent and accurate depth maps from frame sequences remains a formidable challenge. Existing approaches often estimate depth maps for individual frames in isolation, neglecting the rich geometric and temporal coherence present across frames. Consequently, this oversight leads to temporally inconsistent outputs, resulting in noticeable temporal flickering artifacts. In response, this paper presents TCNet, a Temporal Consistent Network for self-supervised monocular depth estimation. Specifically, we propose an Inter-frame Temporal Fusion (ITF) module to emphasize the influence of preceding images on the depth estimation of the current frame. The Temporal Consistency Loss (TCL) is proposed to leverage the temporal constraints between the depth maps of adjacent frames. Besides, TCNet can also be applied to both single-frame and multi-frame scenarios during inference. Experimental evaluations on the KITTI dataset demonstrate that our method surpasses state-of-the-art depth estimation methods in accuracy and temporal consistency. Our code will be made public.

ECAI Conference 2023 Conference Paper

Grafting Fine-Tuning and Reinforcement Learning for Empathetic Emotion Elicitation in Dialog Generation

• Ying Zhu
• Bo Wang 0011
• Dongming Zhao
• Kun Huang
• Zhuoxuan Jiang
• Ruifang He
• Yuexian Hou

For human-like dialogue systems, it is significant to inject the empathetic ability or elicit the opposite’s positive emotions, while existing studies mostly only focus on either of the above two research lines. In this work, we propose a novel and grafted task named Empathetic Emotion Elicitation Dialog to make a dialog system able to possess both aspects of ability simultaneously. We do not train an empathetic dialog system and an emotion elicitation dialog system separately and then simply concatenate the responses generated by these two systems, which will cause illogical and repetitive responses. Instead, we propose a unified solution: (1) To generate empathetic responses and emotion elicitation responses within the same semantic space, we design a unified framework. (2) The unified framework has three stages which first retrieve the empathetic and emotion elicitation exemplars as external knowledge, then fine-tune the emotion/action prediction on a pre-trained language model to enhance the empathetic ability, and finally model the user feedback by reinforcement learning to enhance the emotion elicitation ability. Experiments show that our method outperforms the baselines in the response generation quality and simultaneously empathizes with the user and elicits their positive emotions.

AAAI Conference 2013 Conference Paper

Tools for Preference Reasoning

• Ying Zhu

YNIMG Journal 2007 Journal Article

Neural basis of cultural influence on self-representation

• Ying Zhu
• Li Zhang
• Jin Fan
• Shihui Han

Culture affects the psychological structure of self and results in two distinct types of self-representation (Western independent self and East Asian interdependent self). However, the neural basis of culture–self interaction remains unknown. We used fMRI to measured brain activity from Western and Chinese subjects who judged personal trait adjectives regarding self, mother or a public person. We found that the medial prefrontal cortex (MPFC) and anterior cingulate cortex (ACC) showed stronger activation in self- than other-judgment conditions for both Chinese and Western subjects. However, relative to other-judgments, mother-judgments activated MPFC in Chinese but not in Western subjects. Our findings suggest that Chinese individuals use MPFC to represent both the self and the mother whereas Westerners use MPFC to represent exclusively the self, providing neuroimaging evidence that culture shapes the functional anatomy of self-representation.