Handling Variable-Dimensional Time Series with Graph Neural Networks

TL;DR

This paper introduces a graph neural network-based architecture that enables neural models to handle multi-sensor time series with varying sensor combinations, facilitating zero-shot transfer learning and improved generalization.

Contribution

The paper presents a novel neural network architecture that uses sensor embeddings and graph neural networks to generalize across different sensor combinations in multivariate time series.

Findings

Outperforms baseline models on activity recognition datasets.

Enables zero-shot transfer to unseen sensor combinations.

Improves robustness in multivariate time series modeling.

Abstract

Several applications of Internet of Things (IoT) technology involve capturing data from multiple sensors resulting in multi-sensor time series. Existing neural networks based approaches for such multi-sensor or multivariate time series modeling assume fixed input dimension or number of sensors. Such approaches can struggle in the practical setting where different instances of the same device or equipment such as mobiles, wearables, engines, etc. come with different combinations of installed sensors. We consider training neural network models from such multi-sensor time series, where the time series have varying input dimensionality owing to availability or installation of a different subset of sensors at each source of time series. We propose a novel neural network architecture suitable for zero-shot transfer learning allowing robust inference for multivariate time series with previously unseen combination of available dimensions or sensors at test time. Such a combinatorial generalization is achieved by conditioning the layers of a core neural network-based time series model with a "conditioning vector" that carries information of the available combination of sensors for each time series. This conditioning vector is obtained by summarizing the set of learned "sensor embedding vectors" corresponding to the available sensors in a time series via a graph neural network. We evaluate the proposed approach on publicly available activity recognition and equipment prognostics datasets, and show that the proposed approach allows for better generalization in comparison to a deep gated recurrent neural network baseline.