Model Autophagy
The phenomenon in which a model is repeatedly trained on data it has itself generated, causing errors and biases to be amplified iteratively and weakening the model's generalization ability and diversity.
Core Facts
Timeline (last 90 days)
A 2023 study by Oxford University and other institutions showed that training AI models on data containing large amounts of AI-generated content causes them to lose the ability to capture the tails of the data distribution, leading to homogenization of outputs and quality degradation, known as model autophagy.
An RSS database consisting primarily of human-original content can serve as a high-quality corpus source for AI model training, avoiding the model cannibalization problem.
All Facts (2)
A 2023 study by Oxford University and other institutions showed that training AI models on data containing large amounts of AI-generated content causes them to lose the ability to capture the tails of the data distribution, leading to homogenization of outputs and quality degradation, known as model autophagy.
65%UnverifiedAn RSS database consisting primarily of human-original content can serve as a high-quality corpus source for AI model training, avoiding the model cannibalization problem.
50%