Timezone: »

 
Poster
NAOMI: Non-Autoregressive Multiresolution Sequence Imputation
Yukai Liu · Rose Yu · Stephan Zheng · Eric Zhan · Yisong Yue

Wed Dec 11 05:00 PM -- 07:00 PM (PST) @ East Exhibition Hall B + C #147
in Deep Learning -- Generative Models »

Missing value imputation is a fundamental problem in spatiotemporal modeling, from motion tracking to the dynamics of physical systems. Deep autoregressive models suffer from error propagation which becomes catastrophic for imputing long-range sequences. In this paper, we take a non-autoregressive approach and propose a novel deep generative model: Non-AutOregressive Multiresolution Imputation (NAOMI) to impute long-range sequences given arbitrary missing patterns. NAOMI exploits the multiresolution structure of spatiotemporal data and decodes recursively from coarse to fine-grained resolutions using a divide-and-conquer strategy. We further enhance our model with adversarial training. When evaluated extensively on benchmark datasets from systems of both deterministic and stochastic dynamics. NAOMI demonstrates significant improvement in imputation accuracy (reducing average prediction error by 60% compared to autoregressive counterparts) and generalization for long range sequences.

Keywords: Applications -> Time Series Analysis · Algorithms -> Missing Data; Applications -> Sustainability; Deep Learning -> Adversarial Networks; Deep Learning -> Efficient T

Author Information

Yukai Liu (Caltech)
Rose Yu (Northeastern University)
Stephan Zheng (Salesforce)
Eric Zhan (Caltech)
Yisong Yue (Caltech)

More from the Same Authors