Domain Specific Augmentation

Domain-specific data augmentation (DSA) enhances machine learning model performance by generating synthetic training data tailored to the characteristics of a particular data type (e.g., images, time series, tabular data). Current research focuses on developing DSA methods that improve model generalization to unseen data, including exploring techniques like manifold-based sampling, attention mechanisms, and class-conditioned data corruption, often within contrastive learning frameworks. These advancements are significant because robust DSA strategies are crucial for improving the accuracy and reliability of machine learning models across diverse applications, particularly in scenarios with limited labeled data or significant domain shifts.

Papers

June 16, 2024

First-Order Manifold Data Augmentation for Regression Learning
Ilya Kaufman, Omri Azencot
Data Augmentation Distribution Generalization Domain Agnostic Domain Specific Augmentation Regression via Classification Manifold Augmentation

May 23, 2024

Domain-specific augmentations with resolution agnostic self-attention mechanism improves choroid segmentation in optical coherence tomography images
Jamie Burke, Justin Engelmann, Charlene Hamid, Diana Moukaddem, Dan Pugh, Neeraj Dhaun, Amos Storkey, Niall Strang, Stuart King, Tom MacGillivray, Miguel O. Bernabeu, Ian J. C. MacCormick
Self Attention Optical Coherence Tomography Eye View Conjunct ReSolution Domain Specific Augmentation Choroid Segmentation Sd Oct

April 26, 2024

Tabular Data Contrastive Learning via Class-Conditioned and Feature-Correlation Based Augmentation
Wei Cui, Rasa Hosseinzadeh, Junwei Ma, Tongzi Wu, Yi Sui, Keyvan Golestan
Contrastive Learning Pseudo Labeling Feature Augmentation Tabular Learning Tabular Data Classification Domain Specific Augmentation Tabular Domain

February 22, 2024

Self-Guided Masked Autoencoders for Domain-Agnostic Self-Supervised Learning
Johnathan Xie, Yoonho Lee, Annie S. Chen, Chelsea Finn
Self Supervised Learning Self Supervised Masked Autoencoders Multi Domain Training Domain Specific Augmentation

November 5, 2023

Architecture Matters: Uncovering Implicit Mechanisms in Graph Contrastive Learning
Xiaojun Guo, Yifei Wang, Zeming Wei, Yisen Wang
Contrastive Learning Graph Learning Graph Classification Architecture Design Visual Representation Learning Domain Specific Augmentation

September 16, 2023

Enhancing Visual Perception in Novel Environments via Incremental Data Augmentation Based on Style Transfer
Abhibha Gupta, Rully Agus Hendrawan, Mansur Arief
Generative Model Data Augmentation Style Transfer Visual Perception New Environment Variational Encoder Augmented Dataset Domain Specific Augmentation

May 22, 2023

Distilling Robustness into Natural Language Inference Models with Domain-Targeted Augmentation
Joe Stacey, Marek Rei
Knowledge Distillation Natural Language Inference Robustness Evaluation Distribution Sample Low Temperature Distillation Domain Specific Augmentation

March 22, 2023

On Domain-Specific Pre-Training for Effective Semantic Perception in Agricultural Robotics
Gianmarco Roggiolani, Federico Magistri, Tiziano Guadagnino, Jan Weyler, Giorgio Grisetti, Cyrill Stachniss, Jens Behley
Semantic Segmentation Domain Specific Pre Trained Network Agricultural Robotics Domain Specific Augmentation Leaf Instance Segmentation Semantic Perception

February 23, 2023

Out-of-Domain Robustness via Targeted Augmentations
Irena Gao, Shiori Sagawa, Pang Wei Koh, Tatsunori Hashimoto, Percy Liang
Data Augmentation Unseen Domain Domain Agnostic Domain Specific Augmentation Cross Domain Robustness

January 19, 2023

Foresee What You Will Learn: Data Augmentation for Domain Generalization in Non-stationary Environment
Qiuhao Zeng, Wei Wang, Fan Zhou, Charles Ling, Boyu Wang
Data Augmentation Domain Generalization Unseen Domain Non Stationary Environment Domain Specific Augmentation Domain Transformer

May 30, 2022

Your Contrastive Learning Is Secretly Doing Stochastic Neighbor Embedding
Tianyang Hu, Zhili Liu, Fengwei Zhou, Wenjia Wang, Weiran Huang
Contrastive Learning Self Supervised Contrastive Learning Pairwise Similarity Stochastic Neighbor Domain Specific Augmentation

Domain Specific Augmentation

Papers

First-Order Manifold Data Augmentation for Regression Learning

Domain-specific augmentations with resolution agnostic self-attention mechanism improves choroid segmentation in optical coherence tomography images

Tabular Data Contrastive Learning via Class-Conditioned and Feature-Correlation Based Augmentation

Self-Guided Masked Autoencoders for Domain-Agnostic Self-Supervised Learning

Architecture Matters: Uncovering Implicit Mechanisms in Graph Contrastive Learning

Enhancing Visual Perception in Novel Environments via Incremental Data Augmentation Based on Style Transfer

Distilling Robustness into Natural Language Inference Models with Domain-Targeted Augmentation

On Domain-Specific Pre-Training for Effective Semantic Perception in Agricultural Robotics

Out-of-Domain Robustness via Targeted Augmentations

Foresee What You Will Learn: Data Augmentation for Domain Generalization in Non-stationary Environment

Your Contrastive Learning Is Secretly Doing Stochastic Neighbor Embedding