A5102709279- Advanced Graph Neural Networks
- Anomaly Detection Techniques and Applications
- Hydrology and Watershed Management Studies
- Topic Modeling
- Advanced Text Analysis Techniques
- Complex Network Analysis Techniques
- Flood Risk Assessment and Management
- Software System Performance and Reliability
- Text and Document Classification Technologies
- Cryospheric studies and observations
Henan University
2025
Tianjin Normal University
2024-2025
Chinese Institute for Brain Research
2025
Beijing Normal University
2025
Abstract Sparse precipitation data in karst catchments challenge hydrologic models to accurately capture the spatial and temporal relationships between spring discharge, hindering robust predictions. This study addresses this issue by employing a coupled deep learning model that integrates variation autoencoder (VAE) for augmenting long short‐term memory (LSTM) network discharge prediction. The VAE contributes generating synthetic through an encoding‐decoding process. process generalizes...
Graph anomaly detection is critical in domains such as healthcare and economics, where identifying deviations can prevent substantial losses. Existing unsupervised approaches strive to learn a single model capable of detecting both attribute structural anomalies. However, they confront the tug-of-war problem between two distinct types anomalies, resulting suboptimal performance. This work presents TripleAD, mutual distillation-based triple-channel graph framework. It includes three...
How world knowledge is stored in the human brain a central question cognitive neuroscience. Object effects have been commonly observed higher-order sensory association cortices, with role of language being highly debated. Using object color as test case, we investigated whether communication system plays necessary neural representation visual cortex and corresponding behaviors, combining diffusion imaging (measuring white-matter structural integrity), functional MRI (fMRI; measuring...
Short text can lead to sparse feature representation and classification inaccuracies due noise other issues. To address this, we propose a short model that uses convolutional upsampling enhancement. Our approach involves using multi-scale neural network extract deep features of different dimensions. Secondly, enhance the by convolution obtain more discriminative for downsampling. Finally, use an end-to-end output categories. Experimental validation on public dataset shows our proposed...