Sample Selection Biasedit
How data selection can distort the distribution used for learning or evaluation.
Sample Selection Bias occurs when the data chosen for training or evaluation are not representative of the population or target distribution the model is expected to handle. Selection bias can compound when a model is repeatedly trained on generated or locally filtered data.
Research contextedit
Selection bias is one mechanism behind Synthetic Data failures. Selection bias is not merely a bad dataset label. It is a process: once a subset is preferred, missing modes may receive fewer examples, the model may generate them less often, and the next round of data may become even narrower. In low-resource networked settings, the same mechanism is sharper because rare modes may already be weakly represented before selection starts.
Connection to Qiao's workedit
The ICML 2026 paper When Sample Selection Bias Precipitates Model Collapse studies how local selection behavior can precipitate collapse in recursive synthetic-data training, especially when low-resource verifiers only see fragmented local evidence. This links Qiao's synthetic-data research with AI and networks.