{"ID":2866140,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2509.21473","arxiv_id":"2509.21473","title":"Are Hallucinations Bad Estimations?","abstract":"We formalize hallucinations in generative models as failures to link an estimate to any plausible cause. Under this interpretation, we show that even loss-minimizing optimal estimators still hallucinate. We confirm this with a general high probability lower bound on hallucinate rate for generic data distributions. This reframes hallucination as structural misalignment between loss minimization and human-acceptable outputs, and hence estimation errors induced by miscalibration. Experiments on coin aggregation, open-ended QA, and text-to-image support our theory.","short_abstract":"We formalize hallucinations in generative models as failures to link an estimate to any plausible cause. Under this interpretation, we show that even loss-minimizing optimal estimators still hallucinate. We confirm this with a general high probability lower bound on hallucinate rate for generic data distributions. This...","url_abs":"https://arxiv.org/abs/2509.21473","url_pdf":"https://arxiv.org/pdf/2509.21473v1","authors":"[\"Hude Liu\",\"Jerry Yao-Chieh Hu\",\"Jennifer Yuntong Zhang\",\"Zhao Song\",\"Han Liu\"]","published":"2025-09-25T19:39:09Z","proceeding":"cs.LG","tasks":"[\"cs.LG\",\"cs.AI\",\"cs.CL\",\"cs.CV\",\"stat.ML\"]","methods":"[]","has_code":false}