Awesome-of-Time-Series-Augmentation

A curated list of time series augmentation and related resources.

Please feel free to pull requests or open an issue to add papers.

🔆 Updated 20-01-2022

Table of Contents

• Type of Time-Series-Augmentation

• Time Series Augmentation

  • 2021 Venues
  • 2020 Venues
  • 2019 Venues
  • 2018 Venues
  • 2017 Venues
  • 2016 Venues
  • Previous Venues
  • arXiv

• Awesome Surveys

• Awesome Blogs

• Awesome Videos

Type of Long-tailed Recognition

Type	T	GAN	TA	``	``	``	Other
Explanation	Traditional method	GAN-based method	Text Augmentation				other types

2021

Title	Venue	Type	Code	Star
Improving Stock Price Prediction with GAN-Based Data Augmentation	IJAIDM	GAN	PyTorch(Author)
If You Like It, GAN It—Probabilistic Multivariate Times SeriesForecast with GAN	EP	GAN	PyTorch(Author)
PSA-GAN: Progressive Self Attention GANs for Synthetic Time Series	ICML-W	GAN	PyTorch(Author)

2020

Title	Venue	Type	Code	Star
Conditional GAN for timeseries generation	arXiv	GAN	PyTorch(Author)
Evaluation of Conditional Recurrent Generative Adversarial Networks for Multivariate Time-Series Augmentation	Thesis	GAN	PyTorch(Author)
Data augmentation for time series: traditional vs generative models on capacitive proximity time series	PETRA	GAN	PyTorch(Author)
Generation of Synthetic Financial Time-Series with Generative Adversarial Networks	Thesis	GAN	PyTorch(Author)
Medical Time-Series Data Generation using GenerativeAdversarial Networks		GAN	PyTorch(Author)
End-to-End Differentiable GANs for Text Generation		GAN	PyTorch(Author)

2019

Title	Venue	Type	Code	Star
Time-series Generative Adversarial Networks	NeurIPS	GAN	PyTorch(Author)	247
E2GAN: End-to-End Generative Adversarial Networkfor Multivariate Time Series Imputation	ICLR	GAN	PyTorch(Author)	53
Enriching Financial Datasets with GenerativeAdversarial Networks	Thesis	GAN	PyTorch(Author)
MTSS-GAN: Multivariate Time Series SimulationGenerativeAdversarial Networks	SSRN	GAN	PyTorch(Author)

2018

Title	Venue	Type	Code	Star
Multi-Step Time Series Generator for Molecular Dynamics	AAAI	GAN	PyTorch(Author)
T-CGAN: Conditional Generative Adversarial Network for Data Augmentation in Noisy Time Series with Irregular Sampling	arXiv	GAN	PyTorch(Author)
Real-valued (Medical) Time Series Generation with Recurrent Conditional GANs	ICLR	GAN	PyTorch(Author)	512
Generative Adversarial Network for Synthetic TimeSeries Data Generation in Smart Grids	SC	GAN	PyTorch(Author)

2017

Title	Venue	Type	Code	Star
SeqGAN: Sequence Generative Adversarial Nets with Policy Gradient	AAAI	GAN	PyTorch(Author)	1.9k

2016

Title	Venue	Type	Code	Star
Data Augmentation for Time Series Classification using Convolutional Neural Networks	ECML-W	T	PyTorch(Author)

Previous Venues

Title	Venue	Type	Code	Star
Semi-Supervised Time Series Classification	SIGKDD		PyTorch(Author)
Scheduled Sampling for Sequence Prediction with Recurrent Neural Networks	NIPS	``	PyTorch(Author)

arXiv

Title	Date	Type	Code	Star
Tabular GANs for Uneven Distribution	2020.08.10		-
SeismoGen: Seismic Waveform Synthesis UsingGenerative Adversarial Networks	2020.05.02		-
Statistical analysis of Wasserstein GANswith applications to time seriesforecasting	2020.11.05		-
Adaptive Weighting Scheme for Automatic Time-Series Data Augmentation	2021.02.16		-

Awesome Surveys

• Time Series Data Augmentation for Deep Learning: A Survey, 2021 arXiv

• An empirical survey of data augmentation for time series classification with neural networks, 2021 arXiv

• Text Augmentation Paper List

Awesome Blogs

• 面向深度学习的时间序列数据扩充研究综述

• GAN在自然语言处理方面有哪些有趣的文章和应用？

• 生成式对抗网络GAN有哪些最新的发展，可以实际应用到哪些场景中？

• Role of RL in Text Generation by GAN (强化学习在生成对抗网络文本生成中扮演的角色)

• Data Augmentation 技术调研

• Synthetic Time-Series Data: A GAN approach

• Time Series Forecasting with Recurrent Neural Networks

• Generating Synthetic Sequential Data using GANs

Awesome Videos

• ExperienceAI: GAN for Time Series Data Augmentation in Astronomy - Pavlos Protopapas, Harvard U
• Generation of Synthetic Financial Time Series with GANs - Casper Hogenboom
• TadGAN: Time Series Anomaly Detection Using Generative Adversarial Networks

Related Codebase

Vision

• DAGAN
• Data Augmentation GAN in PyTorch
• Data Augmentation optimized for GAN (DAG) - Official implementation
• Data-Efficient GANs with DiffAugment
• StyleGAN2 with adaptive discriminator augmentation (ADA) — Official TensorFlow implementation
• Semantic preserving image-to-image translation
• Resources and Implementations of Generative Adversarial Nets: GAN, DCGAN, WGAN, CGAN, InfoGAN
• InfoGAN-TensorFlow
• InfoGAN-PyTorch
• infoGAN-pytorch
• Progressive InfoGAN

Language

Time-Series

• TSGAN - TimeSeries - GAN
• semisupervised_timeseries_infogan
• Generative Adversarial Nets for Synthetic Time Series Data
• Generative-Adversarial-Networks-for-financial-time-series-generation
• MTSS-GAN: Multivariate Time Series Simulation Generative Adversarial Networks

Related Papers

Vision-GAN

• Conditional Generative Adversarial Nets, 2014 arXiv
• InfoGAN: Interpretable Representation Learning byInformation Maximizing Generative Adversarial Nets, 2016 NIPS
• Progressive Growing of GANs for Improved Quality, Stability, and Variation, 2018 ICLR
• Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks, 2016 ICLR
• f-GAN: Training Generative Neural Samplers using Variational Divergence Minimization, 2016 NIPS
• Wasserstein GAN, 2017 arXiv
• Improved Training of Wasserstein GANs, 2017 NeurIPS
• Multi-marginal Wasserstein GAN, 2019 NeurIPS

Vision-Augmentation

• Differentiable Augmentation for Data-Efficient GAN Training, 2020 NeurIPS
• Training Generative Adversarial Networks with Limited Data, 2020 NeurIPS
• Effective Data Augmentation with Multi-Domain Learning GANs, 2020 AAAI
• On Data Augmentation for GAN Training, 2020 TIP
• Progressive Augmentation of GANs, 2019 NeurIPS
• A survey on Image Data Augmentation for Deep Learning, 2019 JBD
• GAN Augmentation: Augmenting Training Datausing Generative Adversarial Networks, 2018 arXiv
• Data Augmentation Generative Adversarial Networks, 2017 arXiv
• The Effectiveness of Data Augmentation in Image Classification using Deep Learning, 2017 arXiv

Vision-Translation

• Unpaired Image-to-Image Translationusing Cycle-Consistent Adversarial Networks, 2017 ICCV
• Image-Image Domain Adaptation with Preserved Self-Similarity andDomain-Dissimilarity for Person Re-identification, 2018 CVPR
• AugGAN: Cross Domain Adaptation withGAN-based Data Augmentation, 2018 ECCV

---

Language-GAN

• Connecting Generative Adversarial Networks and Actor-Critic Methods, 2016
• Generative Adversarial Text to Image Synthesis, 2016 ICML
• Professor Forcing: A New Algorithm for Training Recurrent Networks, 2016 NIPS
• Generating Text via Adversarial Training, 2016 NIPS-W
• GANS for Sequences of Discrete Elementswith the Gumbel-softmax Distribution, 2016 NIPS-W
• Adver-sarial Learning for Neural Dialogue Generation, 2017 arXiv
• A Minimax Game for Unifying Generative and Discriminative Information Retrieval Models, 2017 SIGIR