A scalable end-to-end Gaussian process adapter for irregularly sampled time series classification

We present a general framework for classification of sparse and irregularly-sampled time series. The properties of such time series can result in substantial uncertainty about the values of the underlying temporal processes, while making the data difficult to deal with using standard classification methods that assume fixed-dimensional feature spaces. To address these challenges, we propose an uncertainty-aware classification framework based on a special computational layer we refer to as the Gaussian process adapter that can connect irregularly sampled time series data to any black-box classifier learnable using gradient descent. We show how to scale up the required computations based on combining the structured kernel interpolation framework and the Lanczos approximation method, and how to discriminatively train the Gaussian process adapter in combination with a number of classifiers end-to-end using backpropagation.

• Publication year

  2016

• Venue

  Neural Information Processing Systems

• Publication date

  2016-06-14

• Fields of study

  Mathematics, Computer Science

• Identifiers
  arXiv 1606.04443
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Classification of Sparse and Irregularly Sampled Time Series with Mixtures of Expected Gaussian Kernels and Random Features

  2015cited by this paper
• Kernel Interpolation for Scalable Structured Gaussian Processes (KISS-GP)

  2015influential reference
• Sequence to Sequence Learning with Neural Networks

  2014cited by this paper
• Preconditioned Krylov Subspace Methods for Sampling Multivariate Gaussian Distributions

  2014cited by this paper
• Auto-Encoding Variational Bayes

  2013cited by this paper
• Gaussian Process Kernels for Pattern Discovery and Extrapolation

  2013cited by this paper
• Unsupervised pattern discovery in electronic health care data using probabilistic clustering models

  2012cited by this paper
• Matrix Computations

  2011cited by this paper
• A restarted Lanczos approximation to functions of a symmetric matrix

  2010cited by this paper
• uWave: Accelerometer-based personalized gesture recognition and its applications

  2009cited by this paper
• The block Lanczos method for computing eigenvalues

  2007cited by this paper
• Learning methods for generic object recognition with invariance to pose and lighting

  2004cited by this paper
• POPULATION TIME SERIES: PROCESS VARIABILITY, OBSERVATION ERRORS, MISSING VALUES, LAGS, AND HIDDEN STATES

  2004cited by this paper
• Iterative methods for sparse linear systems

  2003cited by this paper
• Retooling the method of block conjugate gradients.

  2001cited by this paper
• The Lomb-Scargle Periodogram in Biological Rhythm Research: Analysis of Incomplete and Unequally Spaced Time-Series

  1999cited by this paper
• Spectrum: spectral analysis of unevenly spaced paleoclimatic time series

  1997cited by this paper
• A block conjugate gradient method applied to linear systems with multiple right-hand sides

  1995cited by this paper
• Solution of the matrix equation AX-XB=C

  1986influential reference
• A Schur method for the square root of a matrix

  1983cited by this paper
• Studies in astronomical time series analysis. II - Statistical aspects of spectral analysis of unevenly spaced data

  1982cited by this paper
• The Symmetric Eigenvalue Problem

  1981cited by this paper
• Cubic convolution interpolation for digital image processing

  1981cited by this paper
• On the Rates of Convergence of the Lanczos and the Block-Lanczos Methods

  1980cited by this paper
• Algorithm 432 [C2]: Solution of the matrix equation AX + XB = C [F4]

  1972cited by this paper

• Generalized Signature Method for Multivariate Time Series: A Data-Driven Framework for Feature Extraction

  2025cites this paper
• SFC: a time series decomposition attention network with continuous nature for time series analysis

  2025cites this paper
• MBLSTM is a contextual interaction refined method for time series prediction

  2025cites this paper
• Iterative Reservoir Computing Networks for Reconstructing Irregular Time Series

  2025cites this paper
• Adapting to the stream: an instance-attention GNN method for irregular multivariate time series data

  2025cites this paper
• Revolutionizing sepsis diagnosis using machine learning and deep learning models: a systematic literature review

  2025cites this paper
• A Survey of Forecasting Methods for Irregular Time Series

  2025cites this paper
• Advancing Precision Oncology Through Modeling of Longitudinal and Multimodal Data

  2025cites this paper
• Augmenting Ground-Level PM2.5 Prediction via Kriging-Based Pseudo-Label Generation

  2024cites this paper
• Using Graphs to Perform Effective Sensor-Based Human Activity Recognition in Smart Homes

  2024cites this paper
• Multi-view representation learning for tabular data integration using inter-feature relationships

  2024cites this paper
• Probabilistic Forecasting of Irregular Time Series via Conditional Flows

  2024cites this paper
• TIMBA: Time series Imputation with Bi-directional Mamba Blocks and Diffusion models

  2024cites this paper
• Data Streams: Investigating Data Structures for Multivariate Asynchronous Time Series Prediction Problems

  2023cites this paper
• Learning to Decouple Complex Systems

  2023cites this paper
• Selective Imputation for Multivariate Time Series Datasets With Missing Values

  2023cites this paper
• Dynamic Graph Attention: Unraveling Spatio-Temporal Synchrony in EEG Data

  2023cites this paper
• Comparing artificial intelligence strategies for early sepsis detection in the ICU: an experimental study

  2023cites this paper
• Know Thy Neighbors: A Graph Based Approach for Effective Sensor-Based Human Activity Recognition in Smart Homes

  2023cites this paper
• Self-attention with temporal prior: can we learn more from the arrow of time?

  2023cites this paper
• On the Generalization Capacities of Neural Controlled Differential Equations

  2023cites this paper
• On the Generalization and Approximation Capacities of Neural Controlled Differential Equations

  2023cites this paper
• Finding Short Signals in Long Irregular Time Series with Continuous-Time Attention Policy Networks

  2023cites this paper
• Automated Model Inference for Gaussian Processes: An Overview of State-of-the-Art Methods and Algorithms

  2022cites this paper
• Tripletformer for Probabilistic Interpolation of Irregularly sampled Time Series

  2022cites this paper
• Kernel Multimodal Continuous Attention

  2022cites this paper
• A posteriori error bounds for the block-Lanczos method for matrix function approximation

  2022cites this paper
• Tripletformer for Probabilistic Interpolation of Asynchronous Time Series

  2022cites this paper
• Liquid Structural State-Space Models

  2022cites this paper
• Stop&Hop: Early Classification of Irregular Time Series

  2022cites this paper
• Sequence Prediction under Missing Data: An RNN Approach without Imputation

  2022cites this paper
• Deep Learning to Jointly Schema Match, Impute, and Transform Databases

  2022cites this paper
• Heteroscedastic Temporal Variational Autoencoder For Irregularly Sampled Time Series

  2022cites this paper
• Experimental Comparison and Survey of Twelve Time Series Anomaly Detection Algorithms

  2021cites this paper
• Finding new edges: systems approaches to MTOR signaling

  2021cites this paper
• Neural SDEs as Infinite-Dimensional GANs

  2021cites this paper
• Causal inference for time series analysis: problems, methods and evaluation

  2021cites this paper
• Learning to Simulate on Sparse Trajectory Data

  2021cites this paper
• The Hintons in your Neural Network: a Quantum Field Theory View of Deep Learning

  2021influential citation
• Self-supervised Transformer for Multivariate Clinical Time-Series with Missing Values

  2021cites this paper
• PIETS: Parallelised Irregularity Encoders for Forecasting with Heterogeneous Time-Series

  2021cites this paper
• Graph-Guided Network for Irregularly Sampled Multivariate Time Series

  2021cites this paper
• Kernel Deformed Exponential Families for Sparse Continuous Attention

  2021cites this paper
• A Generalised Signature Method for Multivariate Time Series Feature Extraction

  2021cites this paper
• Self-Supervised Transformer for Sparse and Irregularly Sampled Multivariate Clinical Time-Series

  2021cites this paper
• Heteroscedastic Temporal Variational Autoencoder for Irregular Time Series.

  2021cites this paper
• A Hybrid Approach for Modeling Type 2 Diabetes Mellitus Progression

  2020cites this paper
• Early Prediction of Sepsis in the ICU Using Machine Learning: A Systematic Review

  2020cites this paper
• Learning Long-Term Dependencies in Irregularly-Sampled Time Series

  2020cites this paper
• Path Imputation Strategies for Signature Models

  2020influential citation
• Generalised Interpretable Shapelets for Irregular Time Series

  2020cites this paper
• A Generalised Signature Method for Time Series

  2020cites this paper
• Multi-Time Attention Networks for Irregularly Sampled Time Series

  2020cites this paper
• Predicting Timing of Surgical Intervention Using Recurrent Neural Network for Necrotizing Pancreatitis

  2020cites this paper
• Large-scale Retrieval of Bayesian Machine Learning Models for Time Series Data via Gaussian Processes

  2020influential citation
• A Review of Deep Learning Methods for Irregularly Sampled Medical Time Series Data

  2020cites this paper
• Probabilistic Numeric Convolutional Neural Networks

  2020influential citation
• A Survey on Principles, Models and Methods for Learning from Irregularly Sampled Time Series: From Discretization to Attention and Invariance

  2020cites this paper
• Automatic Gaussian Process Model Retrieval for Big Data

  2020cites this paper
• Learning from Irregularly-Sampled Time Series

  2020cites this paper
• Neural Controlled Differential Equations for Irregular Time Series

  2020cites this paper
• Forecasting in multivariate irregularly sampled time series with missing values

  2020cites this paper
• Integrating Physiological Time Series and Clinical Notes with Deep Learning for Improved ICU Mortality Prediction

  2020cites this paper
• Progressive Growing of Neural ODEs

  2020cites this paper
• Unsupervised Online Anomaly Detection On Irregularly Sampled Or Missing Valued Time-Series Data Using LSTM Networks

  2020cites this paper
• Early Recognition of Sepsis with Gaussian Process Temporal Convolutional Networks and Dynamic Time Warping

  2019cites this paper
• Temporal Convolutional Networks and Dynamic Time Warping can Drastically Improve the Early Prediction of Sepsis

  2019influential citation
• Dynamic Measurement Scheduling for Event Forecasting using Deep RL

  2019cites this paper
• Deep Prototypical Networks for Imbalanced Time Series Classification under Data Scarcity

  2019cites this paper
• Nonstationary multivariate Gaussian processes for electronic health records

  2019cites this paper
• MGP-AttTCN: An interpretable machine learning model for the prediction of sepsis

  2019cites this paper
• Set Functions for Time Series

  2019cites this paper
• Multi-Task Gaussian Processes and Dilated Convolutional Networks for Reconstruction of Reproductive Hormonal Dynamics

  2019cites this paper
• Dynamic Modeling and Forecasting of Time-evolving Data Streams

  2019cites this paper
• A Robust Two-Sample Test for Time Series data

  2019cites this paper
• Age Time Invariant Features Forecasting Model Time Variant Features ( t = 3 )

  2019cites this paper
• Multi-resolution Networks For Flexible Irregular Time Series Modeling (Multi-FIT)

  2019cites this paper
• Interpolation-Prediction Networks for Irregularly Sampled Time Series

  2019influential citation
• Gaussian Process-Based Models for Clinical Time Series in Healthcare

  2018cites this paper
• Using Large Ensembles of Control Variates for Variational Inference

  2018cites this paper
• Precision and Recall for Time Series

  2018cites this paper
• Unified recurrent neural network for many feature types

  2018cites this paper
• Human Motion Generation via Cross-Space Constrained Sampling

  2018cites this paper
• Hierarchical Gaussian Processes model for multi-task learning

  2018cites this paper
• Deep Learning: Neural Network Applications and Architectures in Sequence and Image Classification

  2018cites this paper
• Modeling Irregularly Sampled Clinical Time Series

  2018cites this paper
• Learning to Detect Sepsis with a Multitask Gaussian Process RNN Classifier

  2017influential citation
• Real-Time Sepsis Prediction using an End-to-End Multi Task Gaussian Process RNN Classifier

  2017cites this paper
• Scalable Joint Models for Reliable Uncertainty-Aware Event Prediction

  2017cites this paper
• Targeting EEG/LFP Synchrony with Neural Nets

  2017influential citation
• An Improved Multi-Output Gaussian Process RNN with Real-Time Validation for Early Sepsis Detection

  2017influential citation
• Autoregressive Convolutional Neural Networks for Asynchronous Time Series

  2017cites this paper
• Active Search with Complex Actions and Rewards

  2014cites this paper
• Knowledge-Based Systems

  year unknowncites this paper
• P RO FIT I : P ROBABILISTIC F ORECASTING OF I RREGULAR T IME S ERIES VIA C ONDITIONAL F LOWS

  year unknowncites this paper