Real-valued (Medical) Time Series Generation with Recurrent Conditional GANs

Abstract

Generative Adversarial Networks (GANs) have shown remarkable success as a framework for training models to produce realistic-looking data. In this work, we propose a Recurrent GAN (RGAN) and Recurrent Conditional GAN (RCGAN) to produce realistic real-valued multi-dimensional time series, with an emphasis on their application to medical data. RGANs make use of recurrent neural networks in the generator and the discriminator. In the case of RCGANs, both of these RNNs are conditioned on auxiliary information. We demonstrate our models in a set of toy datasets, where we show visually and quantitatively (using sample likelihood and maximum mean discrepancy) that they can successfully generate realistic time-series. We also describe novel evaluation methods for GANs, where we generate a synthetic labelled training dataset, and evaluate on a real test set the performance of a model trained on the synthetic data, and vice-versa. We illustrate with these metrics that RCGANs can generate time-series data useful for supervised training, with only minor degradation in performance on real test data. This is demonstrated on digit classification from 'serialised' MNIST and by training an early warning system on a medical dataset of 17,000 patients from an intensive care unit. We further discuss and analyse the privacy concerns that may arise when using RCGANs to generate realistic synthetic medical time series data.

arXiv comment13 pages, 4 figures, 3 tables (update with differential privacy)

References in corpus (11)

• Scikit-learn: Machine Learning in Python
• Conditional Generative Adversarial Nets
• Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks
• Improved Training of Wasserstein GANs
• Improved Techniques for Training GANs
• A Simple Way to Initialize Recurrent Networks of Rectified Linear Units
• C-RNN-GAN: Continuous recurrent neural networks with adversarial training
• SeqGAN: Sequence Generative Adversarial Nets with Policy Gradient
• A Kernel Method for the Two-Sample Problem
• Adversarial Learning for Neural Dialogue Generation
• Variational Recurrent Auto-Encoders

Cited by in corpus (79)

• An Empirical Survey of Data Augmentation for Time Series Classification with Neural Networks
• Generative Adversarial Networks in Computer Vision: A Survey and Taxonomy
• A Review on Generative Adversarial Networks: Algorithms, Theory, and Applications
• Anomaly Detection with Generative Adversarial Networks for Multivariate Time Series
• Synthesizing Tabular Data using Generative Adversarial Networks
• Using GANs for Sharing Networked Time Series Data: Challenges, Initial Promise, and Open Questions
• Classification Accuracy Score for Conditional Generative Models
• Generative Adversarial Networks for Spatio-temporal Data: A Survey
• Probabilistic Forecasting of Sensory Data with Generative Adversarial Networks - ForGAN
• Stabilizing Generative Adversarial Networks: A Survey
• Model-based Deep Reinforcement Learning for Dynamic Portfolio Optimization
• Conditional GAN for timeseries generation
• Feedback GAN (FBGAN) for DNA: a Novel Feedback-Loop Architecture for Optimizing Protein Functions
• A Deep Learning Framework for Generation and Analysis of Driving Scenario Trajectories
• Generative adversarial networks in time series: A survey and taxonomy
• Generative Models for Effective ML on Private, Decentralized Datasets
• Conditional Sig-Wasserstein GANs for Time Series Generation
• COT-GAN: Generating Sequential Data via Causal Optimal Transport
• Generative Adversarial Networks for Electronic Health Records: A Framework for Exploring and Evaluating Methods for Predicting Drug-Induced Laboratory Test Trajectories
• Synthesis of Realistic ECG using Generative Adversarial Networks
• Quick and Easy Time Series Generation with Established Image-based GANs
• Generative Adversarial Networks for Financial Trading Strategies Fine-Tuning and Combination
• Generating Synthetic but Plausible Healthcare Record Datasets
• Generative Adversarial Networks: A Survey Towards Private and Secure Applications
• Synthesizing time-series wound prognosis factors from electronic medical records using generative adversarial networks
• Generalised gravitational burst generation with Generative Adversarial Networks
• TorchGAN: A Flexible Framework for GAN Training and Evaluation
• Measuring the quality of Synthetic data for use in competitions
• A New Multi-vehicle Trajectory Generator to Simulate Vehicle-to-Vehicle Encounters
• A Non-Parametric Test to Detect Data-Copying in Generative Models
• Blood Vessel Geometry Synthesis using Generative Adversarial Networks
• Categorical EHR Imputation with Generative Adversarial Nets
• Interaction-aware Multi-agent Tracking and Probabilistic Behavior Prediction via Adversarial Learning
• TimeVAE: A Variational Auto-Encoder for Multivariate Time Series Generation
• Improved Training with Curriculum GANs
• Generative Models for Simulating Mobility Trajectories
• Time-series Imputation and Prediction with Bi-Directional Generative Adversarial Networks
• Generative Synthesis of Insurance Datasets
• Synthetic Event Time Series Health Data Generation
• Generation of Differentially Private Heterogeneous Electronic Health Records
• Neural SDEs as Infinite-Dimensional GANs
• Deep convolutional generative adversarial networks for traffic data imputation encoding time series as images
• This Car is Mine!: Automobile Theft Countermeasure Leveraging Driver Identification with Generative Adversarial Networks
• Exploring the Vulnerability of Deep Neural Networks: A Study of Parameter Corruption
• Model-Free Reinforcement Learning for Financial Portfolios: A Brief Survey
• Conditional Generative Models for Counterfactual Explanations
• Generative adversarial networks for generation and classification of physical rehabilitation movement episodes
• PAGAN: Portfolio Analysis with Generative Adversarial Networks
• Hide-and-Seek Privacy Challenge
• Towards Better Long-range Time Series Forecasting using Generative Adversarial Networks
• Garment Design with Generative Adversarial Networks
• Visual Evaluation of Generative Adversarial Networks for Time Series Data
• If You Like It, GAN It. Probabilistic Multivariate Times Series Forecast With GAN
• Conditional Generation of Periodic Signals with Fourier-Based Decoder
• Synthetic Epileptic Brain Activities Using Generative Adversarial Networks
• tsBNgen: A Python Library to Generate Time Series Data from an Arbitrary Dynamic Bayesian Network Structure
• Ward2ICU: A Vital Signs Dataset of Inpatients from the General Ward
• GANs with Conditional Independence Graphs: On Subadditivity of Probability Divergences
• Guided-GAN: Adversarial Representation Learning for Activity Recognition with Wearables
• A Novel Framework for Selection of GANs for an Application
• Synthesis of standard 12-lead electrocardiograms using two dimensional generative adversarial network
• Towards Better Adversarial Synthesis of Human Images from Text
• Learning who is in the market from time series: market participant discovery through adversarial calibration of multi-agent simulators
• A Spectral Enabled GAN for Time Series Data Generation
• OCT-GAN: Neural ODE-based Conditional Tabular GANs
• Differentially Private Survival Function Estimation
• Synthetic Data: Opening the data floodgates to enable faster, more directed development of machine learning methods
• Towards Synthetic Multivariate Time Series Generation for Flare Forecasting
• Recurrent Adversarial Service Times
• Generating Multi-type Temporal Sequences to Mitigate Class-imbalanced Problem
• Towards Generating Real-World Time Series Data
• Adversarial Machine Learning-Enabled Anonymization of OpenWiFi Data
• Personality-Driven Gaze Animation with Conditional Generative Adversarial Networks
• Time-Series Regeneration with Convolutional Recurrent Generative Adversarial Network for Remaining Useful Life Estimation
• Adversarial Machine Learning in Text Analysis and Generation
• Generation of Synthetic Electronic Medical Record Text
• Surrogate Monte Carlo
• STEP-GAN: A Step-by-Step Training for Multi Generator GANs with application to Cyber Security in Power Systems
• Toward a Generalization Metric for Deep Generative Models