Your image generator is your new private dataset

Generative diffusion models have emerged as powerful tools to synthetically produce training data, offering potential solutions to data scarcity and reducing labelling costs for downstream supervised deep learning applications. However, existing approaches for synthetic dataset generation face significant limitations: previous methods like Knowledge Recycling rely on label-conditioned generation with models trained from scratch, limiting flexibility and requiring extensive computational resources, while simple class-based conditioning fails to capture the semantic diversity and intra-class variations found in real datasets. Additionally, effectively leveraging text-conditioned image generation for building classifier training sets requires addressing key issues: constructing informative textual prompts, adapting generative models to specific domains, and ensuring robust performance. This paper proposes the Text-Conditioned Knowledge Recycling (TCKR) pipeline to tackle these challenges. TCKR combines dynamic image captioning, parameter-efficient diffusion model fine-tuning, and Generative Knowledge Distillation techniques to create synthetic datasets tailored for image classification. The pipeline is rigorously evaluated on ten diverse image classification benchmarks. The results demonstrate that models trained solely on TCKR-generated data achieve classification accuracies on par with (and in several cases exceeding) models trained on real images. Furthermore, the evaluation reveals that these synthetic-data-trained models exhibit substantially enhanced privacy characteristics: their vulnerability to Membership Inference Attacks is significantly reduced, with the membership inference AUC lowered by 5.49 points on average compared to using real training data, demonstrating a substantial improvement in the performance-privacy trade-off. These findings indicate that high-fidelity synthetic data can effectively replace real data for training classifiers, yielding strong performance whilst simultaneously providing improved privacy protection as a valuable emergent property. The code and trained models are available in the accompanying open-source repository . • Introduces TCKR: text-conditioned diffusion + LoRA + distillation for synthetic data. • Proposes dynamic captioning (BLIP-2) to craft instance-specific prompts for images. • Synthetic-trained classifiers match or surpass real-trained ones on 10 benchmarks. • Synthetic training lowers MIA AUC by up to 8.14 vs models trained on real data. • Increasing synthetic size lifts accuracy but raises MIA risk.