Overview

As an interdisciplinary researcher, my research interests include integrating data statistics into neural network optimization and related applications in time series, i.e., the tasks of generation, prediction, and abnormality detection.

1) Machine Learning(ML) Basics:

Related Topics: First-order Optimization; Convergence Analysis; Domain Adaptation; Predictive Learning; Transformers; Information Theory guided Representation Learning; Physic-inspired Neural Network (NN); Random Matrix Theory; Theoretical Generalization Evaluation.

Example Work I: Leveraging Information Theory Criteria for Deep Domain Adaptation: A Unified Framework and Convergence Analysis. Under review.

Example Work II: Prediction in Prediction: Addressing the Course of Accumulative Prediction Error. Under review.

Example Work III: Conditional Diffusion for Updating Signal-to-Interference Ratio Mapping. Under review.

2) ML (with zero-/few- shot setting) for Autonomous Driving

Related Topics: Open set detection, point cloud recognition/ registration, large-scale point cloud processing, flow estimation, and open-world perception.

Example Work I: Liu, D., Chen, C., Xu, C., Qiu, R. C., & Chu, L^{#}. (2023). Self-supervised point cloud registration with deep versatile descriptors for intelligent driving. IEEE Transactions on Intelligent Transportation Systems, 24(9), 9767-9779.

Example Work II: Liu, D., Chen, C., Xu, C., Cai, Q., Chu, L.^{#}, Wen, F., & Qiu, R. (2022). A robust and reliable point cloud recognition network under rigid transformation. IEEE Transactions on Instrumentation and Measurement, 71, 1-13.

Example Work III: Liu D., Chu, L.^{#}, etc. Multi-Frame Neural Scene Flow.

3) ML (with zero-/few- shot setting) for Sensing and Perception

Example Work I: Chu, L., Pei, L., & Qiu, R. (2021). Ahed: A heterogeneous-domain deep learning model for IoT-enabled smart health with few-labeled EEG data. IEEE Internet of Things Journal, 8(23), 16787-16800.

Example Work II: Chu, L., Qiu, R., He, X., Ling, Z., & Liu, Y. (2017). Massive streaming PMU data modeling and analytics based on multiple high-dimensional covariance test. IEEE Transactions on Big Data, 4(1), 55-64.

Example Work III: Zhang, Y., Xia, S., Chu, L^{#}., Yang, J., Wu, Q., & Pei, L^{#}. (2023). Dynamic Inertial Poser (DynaIP): Part-Based Motion Dynamics Learning for Enhanced Human Pose Estimation with Sparse Inertial Sensors. IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2024.

4) PSG-ML for General Communications and Sensing:

Example Work I: Chu, L., Abdullah A., and Molisch. A. F., "Exploiting semantic localization in highly dynamic wireless networks using deep homoscedastic domain adaptation." arXiv preprint arXiv:2310.07792 (2023), IEEE Transactions on Communications.

Example Work II: Chu, L., Wen, F., Li, L., & Qiu, R. (2019). Efficient nonlinear precoding for massive MIMO downlink systems with 1-bit DACs. IEEE Transactions on Wireless Communications, 18(9), 4213-4224.

Example Work III: Chu, L., Wen, F., & Qiu, R. C. (2019). Eigen-inference precoding for coarsely quantized massive MU-MIMO system with imperfect CSI. IEEE Transactions on Vehicular Technology, 68(9), 8729-8743.