Reference library
AI models collapse when trained on recursively generated data.
Paper summary
Examines repeated training on model-generated data and finds degradation of the represented distribution, including loss of rare content. The analysis and experiments connect recursive replacement of real data with accumulating defects in later model generations.
Relevance to the project
A warning to track source diversity and lineage data if simulated systems or optional language models train from their own outputs. Compare replacement with retained external evidence.
Does not show every mixture of real and synthetic data must collapse.
Source record
- Reference ID
- doi-10-1038-s41586-024-07566-y
- Archived version
- Publisher PDF registered with Crossref
- Summary provenance
- reference-library/papers/doi-10-1038-s41586-024-07566-y/metadata.json · summary
- Review status
- initial orientation summary; not a critical full-paper review
- Reading scope
- Selected abstract/opening and available closing discussion/conclusion passages in the preserved text extract; broader project notes provide context. No reproduction of experiments, exhaustive methods review, or equation/figure audit.