Synthetic Media
New research finds more than a third of freshly published webpages carry signs of AI generation, raising urgent questions about content authenticity, detection, and the future of a synthetic web.
model collapse
Model collapse is the slow-motion degradation that happens when AI models train on AI-generated data. As synthetic media floods the web, future models risk losing diversity, accuracy, and quality in a compounding feedback loop.
model collapse
A new arXiv paper frames model collapse from synthetic data contamination as an epidemiological problem, applying bilayer SIR dynamics to model how AI-generated content spreads through training corpora and degrades future models.
model collapse
As AI-generated content floods the web, models increasingly train on synthetic data, risking model collapse—a degenerative process where outputs lose diversity, accuracy, and connection to reality.
LLM Training
New research introduces SIGMA, a scalable spectral method using eigenvalue analysis to detect model collapse during LLM training before performance degrades catastrophically.
diffusion models
New research reveals how diffusion models suffer 'generative collapse' when trained on synthetic data, with dominated samples disappearing while dominating ones proliferate across generations.