Hanling Yi
Home
Publications
Awards
Projects
Services

Hanling Yi - Projects

A Deep Learning System for Service Fault Prediction in OTN System.

  • Brief Introduction: We formulate the service fault prediction in OTN system as a binary classification problem and propose a deep learning (DL) model to solve this problem. In particular, the proposed DL model consists of a temporal convolution network to extract temporal features from time series KPI data, and a network embedding module to extract spatial features from network topology data. To utilize the unlabeled data, a semi-supervised framework named neural graph machine is integrated in the model. Our proposed DL model is shown to outperform baselines.

A Deep Learning Approach for Battery Charge Time Prediction of Mobile Phone Users.

  • Brief Introduction: By predicting the mobile phone users' battery charge time, we can conduct battery management and improve battery lifetime. We formulate the problem as a sequential prediction problem and propose a GRU-based deep learning model. Our proposed model outperforms traditional ML model such as XGBoost.

A Spatio-temporal GCN Model for Cellular Traffic Prediction.

  • Brief Introduction: The cellular traffic prediction is a spatio-temporal sequential prediction problem. We propose an end-to-end deep learning model that can simultaneously capture the spatial and temporal features. Our DL model consists of graph convolution network (GCN) to extract spatial features and gated recurrent units (GRU) to extract temporal features. Our proposed model can achieve state-of-the-art performance.