Xin Gao
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Xin Gao has a total of 131 co-authors. Here are the top co-authors by number of shared publications:
Bing Xue
School of Intelligent Engineering and Automation, Beijing University of Posts and Telecommunications, Beijing, China
11 shared publicationsBaofeng Li
9 shared publicationsTaizhi Wang
9 shared publicationsXinping Diao
9 shared publicationsFeng Zhai
8 shared publicationsShiyuan Fu
8 shared publications
Latest Publications
- 19 November 2026
- Published by Elsevier in Neurocomputing
A generalized few-shot object detection method with cross-category distillation and contrastive separation space
- 19 November 2026
- Published by Elsevier in Neurocomputing
A generalized few-shot object detection method with cross-category distillation and contrastive separation space
- Open Access
- 1 May 2026
- Published by IOP Publishing in Journal of Physics: Conference Series
- Vol. 3231  (1)
An electric energy curve upsampling method with gated dictionary allocation driven by differential decoupling
Hour-level low-frequency electric energy curves, constrained by acquisition and transmission costs, are insufficient to support fine-grained applications such as load analysis and demand response. Moreover, existing methods lack effective designs for up-sampling monotonic electric energy curves that exhibit high diversity and weak regularities. This paper proposes Conditional Dictionary-based Allocation Predictor (CDAP), a lightweight method tailored for the frequency up-sampling of electric energy curves. CDAP decouples the hourly energy scale and the within-hour fine-grained shape in the differential domain, and constructs an hour-specific logit shape dictionary. A conditional gating network is employed to adaptively mix prototypes, achieving strong representational capacity while keeping the model compact. Additionally, a set of losses, cross-hour smoothing, and entropy regularization is designed to improve training stability. Experiments on one-year real-world data from multiple distribution transformer areas in a southern province of China show that CDAP achieves better overall accuracy, sample-level robustness, and cross-area generalization performance.
- 1 May 2026
- Published by IOP Publishing in Journal of Physics: Conference Series
- Vol. 3231  (1)
Open AccessAn electric energy curve upsampling method with gated dictionary allocation driven by differential decoupling
Hour-level low-frequency electric energy curves, constrained by acquisition and transmission costs, are insufficient to support fine-grained applications such as load analysis and demand response. Moreover, existing methods lack effective designs for up-sampling monotonic electric energy curves that exhibit high diversity and weak regularities. This paper proposes Conditional Dictionary-based Allocation Predictor (CDAP), a lightweight method tailored for the frequency up-sampling of electric energy curves. CDAP decouples the hourly energy scale and the within-hour fine-grained shape in the differential domain, and constructs an hour-specific logit shape dictionary. A conditional gating network is employed to adaptively mix prototypes, achieving strong representational capacity while keeping the model compact. Additionally, a set of losses, cross-hour smoothing, and entropy regularization is designed to improve training stability. Experiments on one-year real-world data from multiple distribution transformer areas in a southern province of China show that CDAP achieves better overall accuracy, sample-level robustness, and cross-area generalization performance.
- 4 April 2026
- Published by Elsevier in Neurocomputing
- Vol. 684
A dual-path fusion network with reconstruction and discrimination for zero-shot multivariate time series anomaly detection
- 4 April 2026
- Published by Elsevier in Neurocomputing
- Vol. 684
A dual-path fusion network with reconstruction and discrimination for zero-shot multivariate time series anomaly detection
- 21 March 2026
- Published by Elsevier in Information Processing & Management
- Vol. 63  (6)
Uncertainty-aware memory-enhanced cross-window dependency learning for robust multivariate time series anomaly detection
- 21 March 2026
- Published by Elsevier in Information Processing & Management
- Vol. 63  (6)
Uncertainty-aware memory-enhanced cross-window dependency learning for robust multivariate time series anomaly detection
- 12 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 13  (11) , 24464-24476
An Adversarial Filtering Framework With TimeâFrequency Cross-Domain Consistency for Multivariate Time-Series Anomaly Detection
Most data collected in real-world scenarios is contaminated, adversely affecting the performance of existing anomaly detection methods that typically rely on clean training data to learn normal patterns. While some approaches have made progress in addressing data contamination, they still face the following challenges: 1) most methods use pixel-level constraints between the contaminated and reconstructed data as the training loss, and the contaminated data is preferentially learned due to the larger error; and 2) current anomaly detection methods based on the frequency domain do not fully utilize the complex spectral characteristics nor the cross-domain consistency of time-series data, which limits the further improvement of the anomaly detection performance. This article proposes a filter-enhanced reconstruction network with domain-adversarial training for multivariate time-series anomaly detection (FERDA), which enables the model to extract normal patterns and high-level semantic information more accurately by training with the error of filtered and reconstructed data. FERDA designs a dual-domain contamination filtering module consisting of the multihead memory module and the complex linear network. The module is trained to learn normal patterns and cross-domain information through adversarial learning of intradomain semantic consistency and interdomain distinguishability. Moreover, a two-stage contamination suppression strategy is proposed. The contamination degree of data is quantified by the difference between pre- and postfiltering, and the impact of contamination is further mitigated by two-stage training with different losses. The proposed method outperforms 23 state-of-the-art baselines in experiments on five benchmark datasets across different domains.
- 12 March 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 13  (11) , 24464-24476
An Adversarial Filtering Framework With TimeâFrequency Cross-Domain Consistency for Multivariate Time-Series Anomaly Detection
Most data collected in real-world scenarios is contaminated, adversely affecting the performance of existing anomaly detection methods that typically rely on clean training data to learn normal patterns. While some approaches have made progress in addressing data contamination, they still face the following challenges: 1) most methods use pixel-level constraints between the contaminated and reconstructed data as the training loss, and the contaminated data is preferentially learned due to the larger error; and 2) current anomaly detection methods based on the frequency domain do not fully utilize the complex spectral characteristics nor the cross-domain consistency of time-series data, which limits the further improvement of the anomaly detection performance. This article proposes a filter-enhanced reconstruction network with domain-adversarial training for multivariate time-series anomaly detection (FERDA), which enables the model to extract normal patterns and high-level semantic information more accurately by training with the error of filtered and reconstructed data. FERDA designs a dual-domain contamination filtering module consisting of the multihead memory module and the complex linear network. The module is trained to learn normal patterns and cross-domain information through adversarial learning of intradomain semantic consistency and interdomain distinguishability. Moreover, a two-stage contamination suppression strategy is proposed. The contamination degree of data is quantified by the difference between pre- and postfiltering, and the impact of contamination is further mitigated by two-stage training with different losses. The proposed method outperforms 23 state-of-the-art baselines in experiments on five benchmark datasets across different domains.
- 25 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Knowledge and Data Engineering
- Vol. 38  (5) , 2508-2524
A Dual Imbalanced Classification Framework With Feature Transfer Guided by Memory Compensation Strategy
Fully mining the differential features of different class samples in overlapping areas is the key and difficult point to improving imbalanced classification performance under complex distribution patterns. Although existing data-level and algorithm-level methods have achieved good results in dealing with overlapping problems, sample generation and classifier training heavily rely on distribution information, and the ability to mine the different information is limited. This paper proposes a dual imbalanced classification framework with feature transfer guided by memory compensation strategy, which enhances the model's ability to mine differential features by constructing a feature space with better inter-class separability. In the traditional classification branch, a feature extraction network maps original samples to feature space and a traditional classifier is used to classify the features. In the compensation classification branch, a feature memory module based on iterative clustering strategy is designed, separately obtaining and saving the correctly classified feature centers of different classes. Moreover, a feature transfer module based on vector combination theory is proposed, combining âpushâ and âpullâ vectors to transfer the misclassified features to the non-overlapping areas corresponding to the same class feature memory module, thereby constructing a feature space with better inter-class separability. Finally, a classification compensation strategy based on feature similarity is designed, integrating the prediction results of the traditional classifier and feature memory module as the final classification results. Experimental results on 50 imbalanced datasets show the proposed method outperforms 28 typical imbalanced classification methods in F1-score and G-mean. Especially on 20 severely overlapping datasets, the performance improvement is more significant.
- 25 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Knowledge and Data Engineering
- Vol. 38  (5) , 2508-2524
A Dual Imbalanced Classification Framework With Feature Transfer Guided by Memory Compensation Strategy
Fully mining the differential features of different class samples in overlapping areas is the key and difficult point to improving imbalanced classification performance under complex distribution patterns. Although existing data-level and algorithm-level methods have achieved good results in dealing with overlapping problems, sample generation and classifier training heavily rely on distribution information, and the ability to mine the different information is limited. This paper proposes a dual imbalanced classification framework with feature transfer guided by memory compensation strategy, which enhances the model's ability to mine differential features by constructing a feature space with better inter-class separability. In the traditional classification branch, a feature extraction network maps original samples to feature space and a traditional classifier is used to classify the features. In the compensation classification branch, a feature memory module based on iterative clustering strategy is designed, separately obtaining and saving the correctly classified feature centers of different classes. Moreover, a feature transfer module based on vector combination theory is proposed, combining âpushâ and âpullâ vectors to transfer the misclassified features to the non-overlapping areas corresponding to the same class feature memory module, thereby constructing a feature space with better inter-class separability. Finally, a classification compensation strategy based on feature similarity is designed, integrating the prediction results of the traditional classifier and feature memory module as the final classification results. Experimental results on 50 imbalanced datasets show the proposed method outperforms 28 typical imbalanced classification methods in F1-score and G-mean. Especially on 20 severely overlapping datasets, the performance improvement is more significant.
- 20 February 2026
- Published by Elsevier in Neural Networks
- Vol. 199, 108731
Memory-guided mask reconstruction with central contrastive learning for robust multivariate time series anomaly detection
- 20 February 2026
- Published by Elsevier in Neural Networks
- Vol. 199, 108731
Memory-guided mask reconstruction with central contrastive learning for robust multivariate time series anomaly detection
- 31 January 2026
- Published by Elsevier in Neural Networks
- Vol. 199, 108652
Adaptive sample repulsion against class-specific counterfactuals for explainable imbalanced classification
- 31 January 2026
- Published by Elsevier in Neural Networks
- Vol. 199, 108652
Adaptive sample repulsion against class-specific counterfactuals for explainable imbalanced classification