NRTSI: Non-Recurrent Time Series Imputation

TL;DR

NRTSI introduces a non-recurrent, permutation-equivariant model for time series imputation that effectively handles irregular sampling, sparsity, and partial observations, achieving state-of-the-art results.

Contribution

The paper proposes a novel non-recurrent, permutation-equivariant approach for time series imputation that improves handling of irregular and sparse data.

Findings

Achieves state-of-the-art performance on multiple benchmarks.

Effectively handles irregularly-sampled and partially observed data.

Supports multiple-mode stochastic imputation.

Abstract

Time series imputation is a fundamental task for understanding time series with missing data. Existing methods either do not directly handle irregularly-sampled data or degrade severely with sparsely observed data. In this work, we reformulate time series as permutation-equivariant sets and propose a novel imputation model NRTSI that does not impose any recurrent structures. Taking advantage of the permutation equivariant formulation, we design a principled and efficient hierarchical imputation procedure. In addition, NRTSI can directly handle irregularly-sampled time series, perform multiple-mode stochastic imputation, and handle data with partially observed dimensions. Empirically, we show that NRTSI achieves state-of-the-art performance across a wide range of time series imputation benchmarks.