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.

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

References in corpus (5)

• Conditional Generative Adversarial Nets
• A Simple Way to Initialize Recurrent Networks of Rectified Linear Units
• C-RNN-GAN: Continuous recurrent neural networks with adversarial training
• A Kernel Method for the Two-Sample Problem
• Variational Recurrent Auto-Encoders

Cited by in corpus (39)

• A Review on Generative Adversarial Networks: Algorithms, Theory, and Applications
• Probabilistic Forecasting of Sensory Data with Generative Adversarial Networks - ForGAN
• Conditional GAN for timeseries generation
• Model-based Deep Reinforcement Learning for Dynamic Portfolio Optimization
• Generative adversarial networks in time series: A survey and taxonomy
• Generative Models for Effective ML on Private, Decentralized Datasets
• Generative Adversarial Networks for Electronic Health Records: A Framework for Exploring and Evaluating Methods for Predicting Drug-Induced Laboratory Test Trajectories
• Generative Adversarial Networks for Financial Trading Strategies Fine-Tuning and Combination
• Generative Adversarial Networks: A Survey Towards Private and Secure Applications
• A Non-Parametric Test to Detect Data-Copying in Generative Models
• Categorical EHR Imputation with Generative Adversarial Nets
• Interaction-aware Multi-agent Tracking and Probabilistic Behavior Prediction via Adversarial Learning
• Time-series Imputation and Prediction with Bi-Directional Generative Adversarial Networks
• Generation of Differentially Private Heterogeneous Electronic Health Records
• Synthetic Event Time Series Health Data Generation
• Model-Free Reinforcement Learning for Financial Portfolios: A Brief Survey
• This Car is Mine!: Automobile Theft Countermeasure Leveraging Driver Identification with Generative Adversarial Networks
• Conditional Generative Models for Counterfactual Explanations
• Generative adversarial networks for generation and classification of physical rehabilitation movement episodes
• Hide-and-Seek Privacy Challenge
• Garment Design with Generative Adversarial Networks
• Visual Evaluation of Generative Adversarial Networks for Time Series Data
• tsBNgen: A Python Library to Generate Time Series Data from an Arbitrary Dynamic Bayesian Network Structure
• Guided-GAN: Adversarial Representation Learning for Activity Recognition with Wearables
• Ward2ICU: A Vital Signs Dataset of Inpatients from the General Ward
• Towards Synthetic Multivariate Time Series Generation for Flare Forecasting
• Towards Better Adversarial Synthesis of Human Images from Text
• Synthetic Data: Opening the data floodgates to enable faster, more directed development of machine learning methods
• Synthesis of standard 12-lead electrocardiograms using two dimensional generative adversarial network
• A Spectral Enabled GAN for Time Series Data Generation
• OCT-GAN: Neural ODE-based Conditional Tabular GANs
• Differentially Private Survival Function Estimation
• Learning who is in the market from time series: market participant discovery through adversarial calibration of multi-agent simulators
• Surrogate Monte Carlo
• Adversarial Machine Learning in Text Analysis and Generation
• Time-Series Regeneration with Convolutional Recurrent Generative Adversarial Network for Remaining Useful Life Estimation
• Personality-Driven Gaze Animation with Conditional Generative Adversarial Networks
• Recurrent Adversarial Service Times
• STEP-GAN: A Step-by-Step Training for Multi Generator GANs with application to Cyber Security in Power Systems