arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2310.13836cs.LGcs.CL
收藏

基础模型的嵌入表示可能检测分布偏移

Foundation Model's Embedded Representations May Detect Distribution Shift

Pacific Northwest National Laboratory(太平洋西北国家实验室);Stony Brook University(石溪大学);University of Washington(华盛顿大学);University of Texas, El Paso(德克萨斯大学埃尔帕索分校)

Max Vargas, Adam Tsou, Andrew Engel, Tony Chiang

AI总结:

本研究通过Sentiment140数据集上的案例,发现预训练基础模型能检测训练与测试集间的分布偏移,并指出线性探测比微调更稳健,有助于模型泛化评估。

摘要:

采样偏差可能导致监督学习任务中训练数据集和测试数据集之间出现分布偏移,从而影响我们理解模型泛化能力。考虑到预训练基础神经网络在迁移学习(TL)任务中的广泛采用——而其行为仍未被充分理解——这一点尤为重要。我们针对Sentiment140数据集上的迁移学习进行了案例研究,并表明许多预训练基础模型对Sentiment140的人工标注测试集$M$与自动标注训练集$P$编码了不同的表示,证实了分布偏移已经发生。我们认为在$P$上训练并在$M$上衡量性能是一种有偏的泛化度量。对预训练GPT-2的实验表明,从$P$中可学习的特征并不能提高(实际上反而损害)在$M$上的性能。对预训练GPT-2表示进行线性探测是稳健的,甚至可能优于整体微调,这暗示了在模型解释中识别训练/测试分割中的分布偏移具有根本重要性。

英文摘要:

Sampling biases can cause distribution shifts between train and test datasets for supervised learning tasks, obscuring our ability to understand the generalization capacity of a model. This is especially important considering the wide adoption of pre-trained foundational neural networks -- whose behavior remains poorly understood -- for transfer learning (TL) tasks. We present a case study for TL on the Sentiment140 dataset and show that many pre-trained foundation models encode different representations of Sentiment140's manually curated test set $M$ from the automatically labeled training set $P$, confirming that a distribution shift has occurred. We argue training on $P$ and measuring performance on $M$ is a biased measure of generalization. Experiments on pre-trained GPT-2 show that the features learnable from $P$ do not improve (and in fact hamper) performance on $M$. Linear probes on pre-trained GPT-2's representations are robust and may even outperform overall fine-tuning, implying a fundamental importance for discerning distribution shift in train/test splits for model interpretation.

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。

↑