mlsanityData to 7 Oct 2026

Synthetic Data for Model Selection

arXiv 2105.00717 ↗First posted 3 May 2021cs.CVcited by 6

Not on the map. Why?

Abstract (arXiv)

Recent breakthroughs in synthetic data generation approaches made it possible to produce highly photorealistic images which are hardly distinguishable from real ones. Furthermore, synthetic generation pipelines have the potential to generate an unlimited number of images. The combination of high photorealism and scale turn synthetic data into a promising candidate for improving various machine learning (ML) pipelines. Thus far, a large body of research in this field has focused on using synthetic images for training, by augmenting and enlarging training data. In contrast to using synthetic data for training, in this work we explore whether synthetic data can be beneficial for model selection. Considering the task of image classification, we demonstrate that when data is scarce, synthetic data can be used to replace the held out validation set, thus allowing to train on a larger dataset. We also introduce a novel method to calibrate the synthetic error estimation to fit that of the real domain. We show that such calibration significantly improves the usefulness of synthetic data for model selection.

Key facts. Cited by 6 papers in the map (data through 7 Oct 2026). Builds on 31 of its 57 references that are in the map. First posted on arXiv 3 May 2021.

Builds on

31 of 57 references.

References · 31

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2020Denoising Diffusion Probabilistic Modelscited by 20,851
  3. 2020Score-Based Generative Modeling through Stochastic Differential Equationscited by 8,964
  4. 2016Densely Connected Convolutional Networkscited by 6,691
  5. 2018A Style-Based Generator Architecture for Generative Adversarial Networkscited by 5,986
  6. 2017Squeeze-and-Excitation Networkscited by 4,763
  7. 2019Analyzing and Improving the Image Quality of StyleGANcited by 3,580
  8. 2020Taming Transformers for High-Resolution Image Synthesiscited by 3,229
  9. 2016Wide Residual Networkscited by 3,057
  10. 2017Progressive Growing of GANs for Improved Quality, Stability, and Variationcited by 2,967
  11. 2018Large Scale GAN Training for High Fidelity Natural Image Synthesiscited by 2,609
  12. 2016Aggregated Residual Transformations for Deep Neural Networkscited by 2,540
19 more
  1. 2017Improved Training of Wasserstein GANscited by 2,223
  2. 2016Identity Mappings in Deep Residual Networkscited by 1,960
  3. 2017Improved Regularization of Convolutional Neural Networks with Cutoutcited by 1,623
  4. 2019Do ImageNet Classifiers Generalize to ImageNet?cited by 1,507
  5. 2020Training Generative Adversarial Networks with Limited Datacited by 1,184
  6. 2019Improved Precision and Recall Metric for Assessing Generative Modelscited by 925
  7. 2017Exploring Generalization in Deep Learningcited by 595
  8. 2018Assessing Generative Models via Precision and Recallcited by 401
  9. 2016Deep Pyramidal Residual Networkscited by 163
  10. 2019Classification Accuracy Score for Conditional Generative Modelscited by 134
  11. 2019Exploring Randomly Wired Neural Networks for Image Recognitioncited by 103
  12. 2020Convolutional Networks with Dense Connectivitycited by 98
  13. 2017Shake-Shake regularizationcited by 87
  14. 2019This dataset does not exist: training models from generated imagescited by 65
  15. 2020Computing the Testing Error without a Testing Setcited by 54
  16. 2020Approximate Cross-validation: Guarantees for Model Assessment and Selectioncited by 14
  17. 2020Approximate Cross-Validation for Structured Modelscited by 9
  18. 2021Ensembles of GANs for synthetic training data generationcited by 6
  19. 2020Leave Zero Out: Towards a No-Cross-Validation Approach for Model Selectioncited by 5

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 1; 2022: 0; 2023: 2; 2024: 1; 2025: 1; 2026: 1.

6 papers.

  1. 2023Utility Theory of Synthetic Data Generationcited by 8
  2. 2023Model Selection of Anomaly Detectors in the Absence of Labeled Validation Datacited by 8
  3. 2024A Study on Training and Developing Large Language Models for Behavior Tree Generationcited by 4
  4. 2021Effective and Privacy preserving Tabular Data Synthesizingcited by 2
  5. 2025Using Synthetic Data to estimate the True Error is theoretically and practically doablecited by 0
  6. 2026When Sample Selection Bias Precipitates Model Collapsecited by 0