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

By Cristóbal Esteban, Stephanie L. Hyland, Gunnar Rätsch. 2017

984 citations recorded by the source corpus.

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 (RNNs) 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, and demonstrate results from differentially private training of the RCGAN.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Generative Models, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Generative Models

In Signal Processing

Citations in the corpus

Privacy-Preserving Generative Deep Neural Networks Support Clinical Data Sharing (2017, background): “A pair of recent preprints have reported generation of synthetic individual participant data via neural networks [4, 5].”
Classification Accuracy Score for Conditional Generative Models (2019, methodology): “Predating [13], [14] used “Train on Synthetic, Test on Real” to measure a recurrent conditional GAN for medical data.”
Generative Models for Effective ML on Private, Decentralized Datasets (2019, background): “Prior work has looked at training GANs with DP in a centralized setting (Torkzadehmahani et al., 2019; Xie et al., 2018; Zhang et al., 2018; Frigerio et al., 2019; Beaulieu-Jones et al., 2018; Esteban et al., 2017).”
A survey of data fusion in smart city applications (2019, methodology): “In addition, we can leverage machine learning approaches such as [243], [244] to generate synthetic datasets with identical data characteristic for study purpose.”
Differentially Private Model Publishing for Deep Learning (2019, methodology): “Several existing DP-SGD implementations [1], [6], [20], [46] use moments accountant(MA) method to track privacy loss, while implementing random shuffling or using its APIs to generate data batches which may violate the random sampling assumption”
Protecting Sensory Data against Sensitive Inferences (2018, background): “For example, to protect health records synthetic medical datasets can be published instead of the real ones using generative models training on sensitive real-world medical datasets [3, 6].”
Deep Self-Organization: Interpretable Discrete Representation Learning on Time Series (2018, background): “Recently, deep generative models have been proposed [10], sometimes even in combination with probabilistic modeling [23].”
Data Augmentation techniques in time series domain: a survey and taxonomy (2022, methodology): “Then they compared the distribution of the data in the two-dimensional space for TimeGAN[55], Recurrent Conditional GAN (GAN)[56], Continuous recurrent GAN (GAN)[58], T-Forcing[59], WaveNet[60] and WaveGAN[61].”
Review of Bridge Structural Health Monitoring Aided by Big Data and Artificial Intelligence: From Condition Assessment to Damage Detection (2020, positive): “In these applications, GANs were used to synthesize medical time series to augment dataset for supervised learning (Esteban et al. 2017), or to learn nonlinear representations for anomaly detection of a machine system in an unsupervised way (Li et al. 2018a).”
Machine Learning for Synthetic Data Generation: a Review (2023, background): “For example, [207] employ synthetic data to train machine learning models and assess their prediction performance on real test data in downstream applications.”
A Review on Generative Adversarial Networks: Algorithms, Theory, and Applications (2020, background): “GANs also have achievements in sequential data such as natural language, music, speech, voice [376], [377], and time series [378]–[381].”
A Survey on Deep Learning for Human Mobility (2020, methodology): “[108] benchmark the performance of RNNs, SeqGAN [226], RGAN [55], and nonparametric copulas to generate synthetic trajectories.”
Generation of Synthetic Electronic Medical Record Text (2018, background): “Hyland proposed a Recurrent (Conditional) GAN to generate real-valued time series in medical application [13].”