正一致训练:无回归模型更新
Positive-Congruent Training: Towards Regression-Free Model Updates
摘要:
在实践中,减少AI系统不同版本行为的不一致性与减少其总体错误同样重要。在图像分类中,样本不一致表现为“负翻转”:新模型错误地预测了旧(参考)模型正确分类的测试样本的输出。正一致(PC)训练旨在降低出错率,同时减少负翻转,从而最大化仅在正预测上与参考模型的一致性,这与模型蒸馏不同。我们提出了一种简单的PC训练方法,焦点蒸馏,它通过为正确分类的样本赋予更多权重来加强与参考模型的一致性。我们还发现,如果参考模型本身可以选择为多个深度神经网络的集合,则可以进一步减少负翻转,而不会影响新模型的准确性。
英文摘要:
Reducing inconsistencies in the behavior of different versions of an AI system can be as important in practice as reducing its overall error. In image classification, sample-wise inconsistencies appear as "negative flips": A new model incorrectly predicts the output for a test sample that was correctly classified by the old (reference) model. Positive-congruent (PC) training aims at reducing error rate while at the same time reducing negative flips, thus maximizing congruency with the reference model only on positive predictions, unlike model distillation. We propose a simple approach for PC training, Focal Distillation, which enforces congruence with the reference model by giving more weights to samples that were correctly classified. We also found that, if the reference model itself can be chosen as an ensemble of multiple deep neural networks, negative flips can be further reduced without affecting the new model's accuracy.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。