| Diverse Intra- and Inter-Domain Activity Style Fusion for Cross-Person Generalizationin Activity Recognition
Junru Zhang, Lang Feng, Zhidan Liu, Yuhan Wu, Yang He, Yabo Dong, Duanqing Xu
ACM SlGKDD International Conference on Knowledge Discovery & Data Mining (KDD) 2024, oral.
TL;DR: Generate novel and unseen time-series instances with mixed activity styles for training to improve domain generalization performance in HAR.
[Paper] |
| Temporal Convolutional Explorer Helps Understand 1D-CNN's Learning Behavior in Time Series Classification from Frequency Domain
Junru Zhang, Lang Feng, Yang He, Yuhan Wu, Yabo Dong
ACM International Conference on Information & Knowledge Management (CIKM) 2023, oral.
TL;DR: Explore the learning behavior of 1D-CNNs in TSC tasks and leverage this finding to correct the suboptimal behavior of 1D-CNNs.
[Paper] [Code][Slide] |
| |
| Adacket: ADAptive Convolutional KErnel Transform for Multivariate Time Series Classification
Junru Zhang, Lang Feng, Haowen Zhang, Yuhan Wu, Yabo Dong
European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML-PKDD) 2023, oral.
TL;DR: An adaptive convolutional kernel transform that models the establishment of 1D-CNN as a sequential decision-making problem using the reinforcement learning.
[Paper] [Code][Slide] [Poster] |
| |
| |
| Federated security tree algorithm for user privacy protection
Junru Zhang, Xiaoyan Zhao, Peiyan Yuan,
Journal of Computer Applications 2023
TL;DR: A federated algorithm for user behavior prediction, ensuring data privacy while enhancing accuracy and computational efficiency.
[Paper] |
| |
| LTCR: Long Temporal Characteristic Reconstruction for Segmentation in Contrastive Learning
Yang He, Yuhan Wu, Junru Zhang, Yabo Dong
European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases (ECML-PKDD) 2024, oral.
TL;DR: A method to enhance contrastive learning for time series by addressing the loss of long temporal characteristics.
|
| Effective LSTMs with seasonal-trend decomposition and adaptive learning and niching-based backtracking search algorithm for time series forecasting
Yuhan Wu, Xiyu Meng, Junru Zhang, Yang He, Joseph A. Romo, Yabo Dong, Dongming Lu
Expert Systems with Applications (ESWA) 2024
TL;DR: A novel hybrid model to automatically optimize the weights and thresholds within each LSTM through an adaptive learning algorithm.
[Paper] |
| |
| Speeding up k-means clustering in high dimensions by pruning unnecessary distance computations
Haowen Zhang, Jing Li, Junru Zhang, Yabo Dong
Knowledge-Based Systems 2023
TL;DR: We propose LBKC, a faster k-means clustering method for high-dimensional data, using innovative lower bounds to maintain accuracy.
[Paper] |
| |
| CLformer: Constraint-based Locality enhanced Transformer for anomaly detection of ancient building structures
Yuhan Wu, Yabo Dong, Wentao Zhu, Junru Zhang, Shijie Liu, Dongming Lu, Nan Zeng, Yinhui Li
Engineering Applications of Artificial Intelligence (EAAI) 2023
TL;DR: Introduce the constraint-based locality enhanced transformer algorithm with multi-scale causal convolution to enhance local feature extraction and advance warning mechanisms for conserving ancient buildings.
[Paper] |
| |
