Machine Learning for Synthetic Data Generation: a Review

Published in

arXiv.org(2023)

External Links:

Generate Graph DownloadPDF

TL;DR

A comprehensive systematic review of existing studies that employ machine learning models for the purpose of generating synthetic data, which explores different machine learning methods, with particular emphasis on neural network architectures and deep generative models.

Abstract

Machine learning heavily relies on data, but real-world applications often encounter various data-related issues. These include data of poor quality, insufficient data points leading to under-fitting of machine learning models, and difficulties in data access due to concerns surrounding privacy, safety, and regulations. In light of these challenges, the concept of synthetic data generation emerges as a promising alternative that allows for data sharing and utilization in ways that real-world data cannot facilitate. This paper presents a comprehensive systematic review of existing studies that employ machine learning models for the purpose of generating synthetic data. The review encompasses various perspectives, starting with the applications of synthetic data generation, spanning computer vision, speech, natural language processing, healthcare, and business domains. Additionally, it explores different machine learning methods, with particular emphasis on neural network architectures and deep generative models. The paper also addresses the crucial aspects of privacy and fairness concerns related to synthetic data generation. Furthermore, this study identifies the challenges and opportunities prevalent in this emerging field, shedding light on the potential avenues for future research. By delving into the intricacies of synthetic data generation, this paper aims to contribute to the advancement of knowledge and inspire further exploration in synthetic data generation.

Authors

Ying-Cheng Lu

1 papers

Huazheng Wang

1 papers

Wenqi Wei

2 papers

References275 items

Differentially Private Fine-Tuning of Diffusion Models

DDN3.0: determining significant rewiring of biological network structure with differential dependency networks

Uncovering LLM-Generated Code: A Zero-Shot Synthetic Code Detector via Code Rewriting

TWIN-GPT: Digital Twins for Clinical Trials via Large Language Model

DP-RDM: Adapting Diffusion Models to Private Domains Without Fine-Tuning

Machine Learning for Synthetic Data Generation: a Review

Published in

arXiv.org(2023)

External Links:

Generate Graph DownloadPDF

TL;DR

Abstract

Authors

Ying-Cheng Lu

1 papers

Huazheng Wang

1 papers

Wenqi Wei

2 papers

References275 items

Differentially Private Fine-Tuning of Diffusion Models

DDN3.0: determining significant rewiring of biological network structure with differential dependency networks

Uncovering LLM-Generated Code: A Zero-Shot Synthetic Code Detector via Code Rewriting

TWIN-GPT: Digital Twins for Clinical Trials via Large Language Model

DP-RDM: Adapting Diffusion Models to Private Domains Without Fine-Tuning

Uncertainty Quantification and Interpretability for Clinical Trial Approval Prediction

Trustworthy Distributed AI Systems: Robustness, Privacy, and Governance

Synthesizing differentially private location traces including co-locations

ABDS: tool suite for analyzing biologically diverse samples

Differentially Private Latent Diffusion Models

Post-processing Private Synthetic Data for Improving Utility on Selected Measures

Meta-hallucinator: Towards Few-Shot Cross-Modality Cardiac Image Segmentation

Boundary crossing: an experimental study of individual perceptions toward AIGC

A Comprehensive Survey of AI-Generated Content (AIGC): A History of Generative AI from GAN to ChatGPT

LDPTrace: Locally Differentially Private Trajectory Synthesis

Extracting Training Data from Diffusion Models

Diffusion Art or Digital Forgery? Investigating Data Replication in Diffusion Models

Reinforced Genetic Algorithm for Structure-based Drug Design

Stutter-TTS: Controlled Synthesis and Improved Recognition of Stuttered Speech

Synthetic Text Generation with Differential Privacy: A Simple and Practical Recipe

Cosbin: cosine score-based iterative normalization of biologically diverse samples

Differentially Private Diffusion Models

Malicious Repositories Detection with Adversarial Heterogeneous Graph Contrastive Learning

Deep Generative Models for Fast Photon Shower Simulation in ATLAS

SynBench: Task-Agnostic Benchmarking of Pretrained Representations using Synthetic Data

PrivTrace: Differentially Private Trajectory Synthesis by Adaptive Markov Model

Artificial intelligence foundation for therapeutic science

DPView: Differentially Private Data Synthesis Through Domain Size Information

Rep2Vec: Repository Embedding via Heterogeneous Graph Adversarial Contrastive Learning

Disentangled Dynamic Heterogeneous Graph Learning for Opioid Overdose Prediction

SIPF: Sampling Method for Inverse Protein Folding

Antibody Complementarity Determining Regions (CDRs) design using Constrained Energy Model

A Multifaceted benchmarking of synthetic electronic health record generation models

Privacy-Preserving Synthetic Data Generation for Recommendation Systems

ACT-NET: Asymmetric Co-Teacher Network for Semi-Supervised Memory-Efficient Medical Image Segmentation

DPGEN: Differentially Private Generative Energy-Guided Network for Natural Image Synthesis

The generation of synthetic data for risk modelling

COT: an efficient and accurate method for detecting marker genes among many subtypes

GeoPointGAN: Synthetic Spatial Data with Local Label Differential Privacy

A Methodology for Controlling Bias and Fairness in Synthetic Data Generation

Video Diffusion Models

MolGenSurvey: A Systematic Survey in Machine Learning Models for Molecule Design

MT-UDA: Towards Unsupervised Cross-modality Medical Image Segmentation with Limited Source Labels

GeoDiff: a Geometric Diffusion Model for Molecular Conformation Generation

Interpretable Molecular Graph Generation via Monotonic Constraints

Deduplicating Training Data Mitigates Privacy Risks in Language Models

Competition-level code generation with AlphaCode

AIM: An Adaptive and Iterative Mechanism for Differentially Private Synthetic Data

Synthetic Data Generation with Differential Privacy via Bayesian Networks

Generative Adversarial Networks

Ethical and social risks of harm from Language Models

Don't Generate Me: Training Differentially Private Generative Models with Sinkhorn Divergence

DECAF: Generating Fair Synthetic Data Using Causally-Aware Generative Networks

Differentially Private Fine-tuning of Language Models

When do GANs replicate? On the choice of dataset size

Fake it till you make it: face analysis in the wild using synthetic data alone

Robin Hood and Matthew Effects: Differential Privacy Has Disparate Impact on Synthetic Data

Differentiable Scaffolding Tree for Molecular Optimization

The Problem of Fairness in Synthetic Healthcare Data

Comparative assessment and novel strategy on methods for imputing proteomics data

Adapting Meta Knowledge with Heterogeneous Information Network for COVID-19 Themed Malicious Repository Detection

Exploring city digital twins as policy tools: A task-based approach to generating synthetic data on urban mobility

Data Synthesis via Differentially Private Markov Random Field

Real-World Trajectory Sharing with Local Differential Privacy

RyanSpeech: A Corpus for Conversational Text-to-Speech Synthesis

An Analysis of the Deployment of Models Trained on Private Tabular Synthetic Data: Unexpected Surprises

SynthASR: Unlocking Synthetic Data for Speech Recognition

Hermite Polynomial Features for Private Data Generation

Exploring the application of synthetic audio in training keyword spotters

Synthetic data in machine learning for medicine and healthcare

Causally Constrained Data Synthesis for Private Data Release

Growing synthetic data through differentially-private vine copulas

Synthetic Data generation using DCGAN for improved traffic sign recognition

DDN2.0: R and Python packages for differential dependency network analysis of biological systems

Can You Fake It Until You Make It?: Impacts of Differentially Private Synthetic Data on Downstream Classification Fairness

Therapeutics Data Commons: Machine Learning Datasets and Tasks for Therapeutics

HINT: Hierarchical interaction network for clinical-trial-outcome predictions

GraphEBM: Molecular Graph Generation with Energy-Based Models

Generating and evaluating cross‐sectional synthetic electronic healthcare data: Preserving data utility and patient privacy

Location Privacy-preserving Mechanisms in Location-based Services

Kamino: Constraint-Aware Differentially Private Data Synthesis

Differentially Private Synthetic Medical Data Generation using Convolutional GANs

Extracting Training Data from Large Language Models

dStyle-GAN: Generative Adversarial Network based on Writing and Photography Styles for Drug Identification in Darknet Markets

DeepPurpose: a deep learning library for drug–target interaction prediction

Score-Based Generative Modeling through Stochastic Differential Equations

Using Synthetic Audio to Improve the Recognition of Out-of-Vocabulary Words in End-to-End Asr Systems

Synthetic Data - Anonymisation Groundhog Day

Generating high-fidelity synthetic patient data for assessing machine learning healthcare software

Community Mitigation: A Data-driven System for COVID-19 Risk Assessment in a Hierarchical Manner

MIMOSA: Multi-constraint Molecule Sampling for Molecule Optimization

A Framework for Evaluating Client Privacy Leakages in Federated Learning

Mace: A flexible framework for membership privacy estimation in generative models

WaveGrad: Estimating Gradients for Waveform Generation

Mixing Real and Synthetic Data to Enhance Neural Network Training - A Review of Current Approaches

$\alpha$-Satellite: An AI-Driven System and Benchmark Datasets for Dynamic COVID-19 Risk Assessment in the United States

NVAE: A Deep Hierarchical Variational Autoencoder

Generating sequential electronic health records using dual adversarial autoencoder

100

Generating synthetic data in finance: opportunities, challenges and pitfalls

101

Denoising Diffusion Probabilistic Models

102

GS-WGAN: A Gradient-Sanitized Approach for Learning Differentially Private Generators

103

Language Models are Few-Shot Learners

104

Generation and evaluation of synthetic patient data

105

A Non-Parametric Test to Detect Data-Copying in Generative Models

106

BLEURT: Learning Robust Metrics for Text Generation

107

Inverting Gradients - How easy is it to break privacy in federated learning?

108

DP-MERF: Differentially Private Mean Embeddings with RandomFeatures for Practical Privacy-preserving Data Generation

109

CorGAN: Correlation-Capturing Convolutional Generative Adversarial Networks for Generating Synthetic Healthcare Records

110

Compressive Privacy Generative Adversarial Network

111

DeepSynth: Three-dimensional nuclear segmentation of biological images using neural networks trained with synthetic data

112

CORE: Automatic Molecule Optimization Using Copy & Refine Strategy

113

Federated Generative Privacy

114

Is this Real?: Generating Synthetic Data that Looks Real

115

PPGAN: Privacy-Preserving Generative Adversarial Network

116

Differentially Private and Utility Preserving Publication of Trajectory Data

117

Synthetic Data for Deep Learning

118

GAN-Leaks: A Taxonomy of Membership Inference Attacks against Generative Models

119

PEARL: Prototype Learning via Rule Learning

120

On the Utility of Synthetic Data: An Empirical Evaluation on Machine Learning Tasks

121

A Survey on Bias and Fairness in Machine Learning

122

DDL: Deep Dictionary Learning for Predictive Phenotyping

123

A Survey on Federated Learning Systems: Vision, Hype and Reality for Data Privacy and Protection

124

A Brief History of Artificial Intelligence: On the Past, Present, and Future of Artificial Intelligence

125

Quant GANs: deep generation of financial time series

126

Empirical Evaluation on Synthetic Data Generation with Generative Adversarial Network

127

G-PATE: Scalable Differentially Private Data Generator via Private Aggregation of Teacher Discriminators

128

Integrated Identification of Disease Specific Pathways Using Multi-omics data

129

Monte Carlo and Reconstruction Membership Inference Attacks against Generative Models

130

Generating Diverse High-Fidelity Images with VQ-VAE-2

131

DP-CGAN: Differentially Private Synthetic Data and Label Generation

132

Deep Leakage from Gradients

134

Strategies for Pre-training Graph Neural Networks

135

Differential Privacy Has Disparate Impact on Model Accuracy

136

Generative Adversarial Networks for Mitigating Biases in Machine Learning Systems

137

Graph Convolutional Networks with EigenPooling

138

Data Augmentation Using GANs

139

Time Series Simulation by Conditional Generative Adversarial Net

140

A New GAN-based End-to-End TTS Training Algorithm

141

Speech synthesis from neural decoding of spoken sentences

142

Data scarcity, robustness and extreme multi-label classification

143

Learning From Synthetic Data for Crowd Counting in the Wild

144

SynSys: A Synthetic Data Generation System for Healthcare Applications

145

The Transactions on Intelligent Vehicles

146

Graph Neural Networks for Social Recommendation

147

GANobfuscator: Mitigating Information Leakage Under GAN via Differential Privacy

148

Demystifying Membership Inference Attacks in Machine Learning as a Service

149

Differentially Private Release of High-Dimensional Datasets using the Gaussian Copula

150

Towards a Deeper Understanding of Adversarial Losses

151

Data-driven detection of subtype-specific and differentially expressed genes

152

Differentially Private Generative Adversarial Networks for Time Series, Continuous, and Discrete Open Data

153

Learning Semantic Segmentation From Synthetic Data: A Geometrically Guided Input-Output Adaptation Approach

154

Differentially Private Data Generative Models

155

Beyond Inferring Class Representatives: User-Level Privacy Leakage From Federated Learning

156

A graph-based genetic algorithm and generative model/Monte Carlo tree search for the exploration of chemical space

157

Optimization of Molecules via Deep Reinforcement Learning

158

Utility-Aware Synthesis of Differentially Private and Attack-Resilient Location Traces

159

CALM: Consistent Adaptive Local Marginal for Marginal Release under Local Differential Privacy

160

Large Scale GAN Training for High Fidelity Natural Image Synthesis

161

PATE-GAN: Generating Synthetic Data with Differential Privacy Guarantees

162

Privacy Preserving Synthetic Data Release Using Deep Learning

163

Diffusion Models

164

Multi-omics Data Integration for Identifying Disease Specific Biological Pathways

165

Natural language generation for electronic health records

166

MolGAN: An implicit generative model for small molecular graphs

167

FairGAN: Fairness-aware Generative Adversarial Networks

168

Exploiting Unintended Feature Leakage in Collaborative Learning

169

Toward Diverse Text Generation with Inverse Reinforcement Learning

170

Understanding disentangling in β-VAE

171

BAGAN: Data Augmentation with Balancing GAN

172

Enhance wound healing monitoring through a thermal imaging based smartphone app

173

Differentially Private Generative Adversarial Network

174

Adversarial Audio Synthesis

175

Junction Tree Variational Autoencoder for Molecular Graph Generation

176

Synthetic data augmentation using GAN for improved liver lesion classification

177

A Flexible Method for Protecting Marketing Data: An Application to Point-of-Sale Data

178

Learning from Synthetic Data: Addressing Domain Shift for Semantic Segmentation

179

POSTER: A Unified Framework of Differentially Private Synthetic Data Release with Generative Adversarial Network

180

Synthetic Data for Social Good

181

Differentially Private Mixture of Generative Neural Networks

182

Plausible Deniability for Privacy-Preserving Data Synthesis

183

Security Techniques for the Electronic Health Records

184

Wasserstein Generative Adversarial Networks

185

Privacy-Preserving Generative Deep Neural Networks Support Clinical Data Sharing

186

Attention is All you Need

187

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

188

Protecting personal trajectories of social media users through differential privacy

189

LOGAN: Membership Inference Attacks Against Generative Models

190

Molecular de-novo design through deep reinforcement learning

191

Affect-LM: A Neural Language Model for Customizable Affective Text Generation

192

Generating Multi-label Discrete Patient Records using Generative Adversarial Networks

193

Deep Models Under the GAN: Information Leakage from Collaborative Deep Learning

194

HoloClean: Holistic Data Repairs with Probabilistic Inference

195

Compressive Privacy: From Information\/Estimation Theory to Machine Learning [Lecture Notes]

196

synthpop: Bespoke Creation of Synthetic Data in R

197

Membership Inference Attacks Against Machine Learning Models

198

Automatic Chemical Design Using a Data-Driven Continuous Representation of Molecules

199

The Synthetic Data Vault

200

SeqGAN: Sequence Generative Adversarial Nets with Policy Gradient

201

WaveNet: A Generative Model for Raw Audio

202

Semi-Supervised Classification with Graph Convolutional Networks

203