Research map / Label-efficient and robust learning / Recognition and long-tail
Bayesian Generative Active Deep Learning
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Condition VAE-ACGAN augmentation on BALD-selected samples rather than augmenting without information-based selection | Active learning with ResNet18 · CIFAR-10 | accuracy: 86.80 ± 0.45 -> 87.63 ± 0.11 | Helped |
| Condition VAE-ACGAN augmentation on BALD-selected samples rather than augmenting without information-based selection | Active learning with ResNet18pa · CIFAR-10 | accuracy: 90.70 ± 0.24 -> 91.13 ± 0.10 | Helped |
| Condition VAE-ACGAN augmentation on BALD-selected samples rather than augmenting without information-based selection | Active learning with ResNet18 · CIFAR-100 | accuracy: 66.50 ± 0.63 -> 68.05 ± 0.17 | Helped |
| Condition VAE-ACGAN augmentation on BALD-selected samples rather than augmenting without information-based selection | Active learning with ResNet18pa · CIFAR-100 | accuracy: 67.79 ± 0.76 -> 69.69 ± 0.13 | Helped |
| Use VAEACGAN-generated augmentation instead of active learning without data augmentation | Active learning with ResNet18 · CIFAR-100 | accuracy: 50.57 ± 0.20 -> 68.05 ± 0.17 | Helped |
| Use VAEACGAN-generated augmentation instead of active learning without data augmentation | Active learning with ResNet18pa · CIFAR-100 | accuracy: 55.82 ± 0.31 -> 69.69 ± 0.13 | Helped |
| Use informative active selection and generated samples rather than partial-training BDA | Bayesian generative active learning · CIFAR-10 | accuracy: 85.08 ± 0.31 -> 87.63 ± 0.11 | Helped |
| Use informative active selection and generated samples rather than partial-training BDA | Bayesian generative active learning · CIFAR-100 | accuracy: 65.76 ± 0.40 -> 68.05 ± 0.17 | Helped |
| Use informative active selection and generated samples rather than partial-training BDA | Bayesian generative active learning · CIFAR-100 | accuracy: 65.79 ± 0.51 -> 69.69 ± 0.13 | Helped |
| Use active selection instead of random sample selection | Bayesian generative active learning · CIFAR-10 | accuracy: 77.29 ± 0.23 -> 87.63 ± 0.11 | Helped |
Closest papers by meaning
- A Bayesian Data Augmentation Approach for Learning Deep Models (NeurIPS 2017 · similarity 0.89)
- Towards Controlled Data Augmentations for Active Learning (ICML 2023 · similarity 0.87)
- Deep Bayesian Active Learning with Image Data (ICML 2017 · similarity 0.86)
- Data-Efficient Augmentation for Training Neural Networks (NeurIPS 2022 · similarity 0.86)
- Generative Active Learning via Bayesian Acquisition for Improving the Efficiency of Synthetic Data (NeurIPS 2026 · similarity 0.86)
- Active Learning Guided by Efficient Surrogate Learners (AAAI 2024 · similarity 0.84)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.