Networked Time Series Imputation via Position-aware Graph Enhanced Variational Autoencoders

Abstract

Multivariate time series (MTS) imputation is a widely studied problem in recent years. Existing methods can be divided into two main groups, including (1) deep recurrent or generative models that primarily focus on time series features, and (2) graph neural networks (GNNs) based models that utilize the topological information from the inherent graph structure of MTS as relational inductive bias for imputation. Nevertheless, these methods either neglect topological information or assume the graph structure is fixed and accurately known. Thus, they fail to fully utilize the graph dynamics for precise imputation in more challenging MTS data such as networked time series (NTS), where the underlying graph is constantly changing and might have missing edges. In this paper, we propose a novel approach to overcome these limitations. First, we define the problem of imputation over NTS which contains missing values in both node time series features and graph structures. Then, we design a new model named PoGeVon which leverages variational autoencoder (VAE) to predict missing values over both node time series features and graph structures. In particular, we propose a new node position embedding based on random walk with restart (RWR) in the encoder with provable higher expressive power compared with message-passing based graph neural networks (GNNs). We further design a decoder with 3-stage predictions from the perspective of multi-task learning to impute missing values in both time series and graph structures reciprocally. Experiment results demonstrate the effectiveness of our model over baselines.

arXiv commentKDD 2023

References in corpus (30)

• Adam: A Method for Stochastic Optimization
• Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling
• The Emerging Field of Signal Processing on Graphs: Extending High-Dimensional Data Analysis to Networks and Other Irregular Domains
• Neural Message Passing for Quantum Chemistry
• The information bottleneck method
• Tutorial on Variational Autoencoders
• Linformer: Self-Attention with Linear Complexity
• GAIN: Missing Data Imputation using Generative Adversarial Nets
• SAITS: Self-Attention-based Imputation for Time Series
• TimesNet: Temporal 2D-Variation Modeling for General Time Series Analysis
• BRITS: Bidirectional Recurrent Imputation for Time Series
• Understanding disentangling in $\beta$-VAE
• HDMI: High-order Deep Multiplex Infomax
• Examining COVID-19 Forecasting using Spatio-Temporal Graph Neural Networks
• Graph Neural Networks for Graphs with Heterophily: A Survey
• Distance Encoding: Design Provably More Powerful Neural Networks for Graph Representation Learning
• Adaptive Universal Generalized PageRank Graph Neural Network
• Variational Autoencoder with Arbitrary Conditioning
• Large Scale Learning on Non-Homophilous Graphs: New Benchmarks and Strong Simple Methods
• Personalized PageRank to a Target Node, Revisited
• Z-GCNETs: Time Zigzags at Graph Convolutional Networks for Time Series Forecasting
• Filling the G_ap_s: Multivariate Time Series Imputation by Graph Neural Networks
• Time Series Data Imputation: A Survey on Deep Learning Approaches
• Network of Tensor Time Series
• HIST: A Graph-based Framework for Stock Trend Forecasting via Mining Concept-Oriented Shared Information
• Equivariant and Stable Positional Encoding for More Powerful Graph Neural Networks
• Modelling, Detrending and Decorrelation of Network Time Series
• Learning to Reconstruct Missing Data from Spatiotemporal Graphs with Sparse Observations
• VAEs in the Presence of Missing Data
• Retrieval Based Time Series Forecasting

Cited by in corpus (6)

• Graph Deep Learning for Time Series Forecasting
• Scalable Transformer for High Dimensional Multivariate Time Series Forecasting
• Boundary-enhanced time series data imputation with long-term dependency diffusion models
• Causality-Aware Spatiotemporal Graph Neural Networks for Spatiotemporal Time Series Imputation
• Automated Contrastive Learning Strategy Search for Time Series
• Modeling Time-evolving Causality over Data Streams