Synthetic Data Augmentation

Synthetic data augmentation generates artificial data to supplement real-world datasets, primarily addressing limitations in data quantity, diversity, and balance crucial for training robust machine learning models. Current research focuses on using generative models like GANs, VAEs, and diffusion models to create realistic synthetic data for various applications, including medical image analysis, autonomous driving, and natural language processing. This technique significantly impacts fields with limited real data availability, improving model performance and generalizability while potentially reducing the cost and effort associated with data collection and annotation.

Papers

November 14, 2022

A deep learning framework to generate realistic population and mobility data
Eren Arkangil, Mehmet Yildirimoglu, Jiwon Kim, Carlo Prato
Deep Learning Framework Mobility Data Travel Demand Synthetic Data Augmentation Synthetic Population Travel Behavior Household Travel Survey

November 10, 2022

Scalable Modular Synthetic Data Generation for Advancing Aerial Autonomy
Mehrnaz Sabet, Praveen Palanisamy, Sakshi Mishra
Synthetic Data Generation Synthetic Data Augmentation Aerial Datasets Large Scale Aerial Datasets

October 29, 2022

Single-Shot Domain Adaptation via Target-Aware Generative Augmentation
Rakshith Subramanyam, Kowshik Thopalli, Spring Berman, Pavan Turaga, Jayaraman J. Thiagarajan
Domain Adaptation Data Augmentation Shot Adaptation Synthetic Data Augmentation One Shot Unsupervised Domain Adaptation

October 7, 2022

Trustworthiness of Laser-Induced Breakdown Spectroscopy Predictions via Simulation-based Synthetic Data Augmentation and Multitask Learning
Riccardo Finotello, Daniel L'Hermite, Celine Quéré, Benjamin Rouge, Mohamed Tamaazousti, Jean-Baptiste Sirven
Multitask Learning Spectral Information Synthetic Data Augmentation Experimental Data Spectral Data Laser Induced Breakdown Spectroscopy

October 6, 2022

Fault Diagnosis using eXplainable AI: a Transfer Learning-based Approach for Rotating Machinery exploiting Augmented Synthetic Data
Lucas Costa Brito, Gian Antonio Susto, Jorge Nei Brito, Marcus Antonio Viana Duarte
Transfer Learning Explainable AI Explainable Artificial Intelligence Learning Based Approach Vibration Signal Synthetic Data Augmentation Industrial Machine

September 14, 2022

DASH: Visual Analytics for Debiasing Image Classification via User-Driven Synthetic Data Augmentation
Bum Chul Kwon, Jungsoo Lee, Chaeyeon Chung, Nyoungwoo Lee, Ho-Jin Choi, Jaegul Choo
Bias Mitigation Image to Image Bias Evaluation Visual Analytics Image Classification Model Synthetic Data Augmentation Data Bias

June 29, 2022

Distilling Model Failures as Directions in Latent Space
Saachi Jain, Hannah Lawrence, Ankur Moitra, Aleksander Madry
Latent Space Interpretable Direction Synthetic Data Augmentation Human Representation Model Failure Error Pattern

March 16, 2022

Sample, Translate, Recombine: Leveraging Audio Alignments for Data Augmentation in End-to-end Speech Translation
Tsz Kin Lam, Shigehiko Schamoni, Stefan Riezler
Data Augmentation Speech Translation Language Pair End to End Speech Translation Source Speech Synthetic Data Augmentation Model Recombination Audio Alignment

December 14, 2021

Improving COVID-19 CXR Detection with Synthetic Data Augmentation
Daniel Schaudt, Christopher Kloth, Christian Spaete, Andreas Hinteregger, Meinrad Beer, Reinhold von Schwerin
Generative Adversarial Network Synthetic Image Synthetic Data Augmentation Synthetic Chest X Ray

November 25, 2021

Generalizing electrocardiogram delineation -- Training convolutional neural networks with synthetic data augmentation
Guillermo Jimenez-Perez, Juan Acosta, Alejandro Alcaine, Oscar Camara
Convolutional Neural Network Electrocardiogram Signal Synthetic Data Augmentation Beat Tracking Electrocardiogram Recording Electrocardiogram Delineation

Synthetic Data Augmentation

Papers

A deep learning framework to generate realistic population and mobility data

Scalable Modular Synthetic Data Generation for Advancing Aerial Autonomy

Single-Shot Domain Adaptation via Target-Aware Generative Augmentation

Trustworthiness of Laser-Induced Breakdown Spectroscopy Predictions via Simulation-based Synthetic Data Augmentation and Multitask Learning

Fault Diagnosis using eXplainable AI: a Transfer Learning-based Approach for Rotating Machinery exploiting Augmented Synthetic Data

DASH: Visual Analytics for Debiasing Image Classification via User-Driven Synthetic Data Augmentation

Distilling Model Failures as Directions in Latent Space

Sample, Translate, Recombine: Leveraging Audio Alignments for Data Augmentation in End-to-end Speech Translation

Improving COVID-19 CXR Detection with Synthetic Data Augmentation

Generalizing electrocardiogram delineation -- Training convolutional neural networks with synthetic data augmentation