synthetic samples
Synthetic samples are artificially generated data points used to augment a dataset. They can help models learn better by providing more training data or by introducing variability, especially when labeled real-world data is scarce.
- Hyperbolic Dataset Distillation
- On the Closed-Form of Flow Matching: Generalization Does Not Arise from Target Stochasticity
- On the Closed-Form of Flow Matching: Generalization Does Not Arise from Target Stochasticity
- SPACE: Noise Contrastive Estimation Stabilizes Self-Play Fine-Tuning for Large Language Models
- Unlocking Dataset Distillation with Diffusion Models
- Valid Inference with Imperfect Synthetic Data