Model kolaps

It is a research finding under particular data and training conditions, not proof that every use of synthetic data will fail.

Klíčové věci

• State the recursive-training conditions.
• Preserve provenance and independent evaluation.
• Inspect rare cases and diversity.

A generator approximates patterns in its training distribution. If a later model is trained mainly on samples from that approximation, errors and missing rare cases can propagate. Repeating the process can narrow what the models represent. The 2024 Nature study investigates this behavior across several model families. The training setup matters. Replacing original data with generated outputs is different from retaining independently collected data while adding selected synthetic examples. Filtering, sampling, objectives, and evaluation can affect outcomes. Avoid treating all synthetic-data strategies as one experiment. Record the provenance and generation process for training material. Keep an independently sourced evaluation set that is not regenerated by the model being assessed. Measure rare categories and diversity as well as common-case accuracy, because loss of coverage can be hidden by an average. When testing synthetic augmentation, compare a real-data baseline, the proposed mixture, and relevant alternatives under the same budget. Report which conditions improved or degraded. A useful conclusion describes the tested setup and uncertainty rather than predicting an inevitable fate for all AI systems.

1. Construct a toy dataset with 90 examples of a common pattern and 10 of a rare pattern.
2. Suppose a generator produces only two rare-pattern examples in its next 100 samples. Training solely on those outputs changes the represented balance.
3. Measure rare-pattern performance against the original held-out data before repeating the cycle.

This invented scenario illustrates a possible mechanism, not the quantitative result of the cited study.