arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2202.09717cs.LG
收藏

理解正则语言学习中的鲁棒泛化

Understanding Robust Generalization in Learning Regular Languages

University of Pennsylvania(宾夕法尼亚大学)

Soham Dan, Osbert Bastani, Dan Roth

摘要:

人类智能的一个关键特征是超越训练分布的能力,例如,解析比过去更长的句子。目前,深度神经网络很难对数据分布中的这种变化进行鲁棒的概括。我们在使用递归神经网络(RNN)学习正则语言的背景下研究鲁棒泛化。我们假设标准的端到端建模策略不能很好地推广到系统的分布偏移,并提出了一个组合策略来解决这个问题。我们比较了一个端到端的战略,将字符串映射到标签的组合策略,预测结构的确定性有限状态自动机(DFA),接受正则语言。我们从理论上证明,组合策略的推广显着优于端到端的策略。在我们的实验中,我们通过一个辅助任务来实现组合策略,其目标是预测DFA在解析字符串时访问的中间状态。我们的实证结果支持我们的假设,表明辅助任务可以实现强大的泛化。有趣的是,端到端RNN的泛化能力明显优于理论下限,这表明它至少能够实现某种程度的鲁棒泛化。

英文摘要:

A key feature of human intelligence is the ability to generalize beyond the training distribution, for instance, parsing longer sentences than seen in the past. Currently, deep neural networks struggle to generalize robustly to such shifts in the data distribution. We study robust generalization in the context of using recurrent neural networks (RNNs) to learn regular languages. We hypothesize that standard end-to-end modeling strategies cannot generalize well to systematic distribution shifts and propose a compositional strategy to address this. We compare an end-to-end strategy that maps strings to labels with a compositional strategy that predicts the structure of the deterministic finite-state automaton (DFA) that accepts the regular language. We theoretically prove that the compositional strategy generalizes significantly better than the end-to-end strategy. In our experiments, we implement the compositional strategy via an auxiliary task where the goal is to predict the intermediate states visited by the DFA when parsing a string. Our empirical results support our hypothesis, showing that auxiliary tasks can enable robust generalization. Interestingly, the end-to-end RNN generalizes significantly better than the theoretical lower bound, suggesting that it is able to achieve at least some degree of robust generalization.

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。

↑