arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2106.07804cs.LGstat.ML
收藏

用规则表示控制神经网络

Controlling Neural Networks with Rule Representations

Google Cloud AI(谷歌云AI)

Sungyong Seo, Sercan O. Arik, Jinsung Yoon, Xiang Zhang, Kihyuk Sohn, Tomas Pfister

摘要:

我们提出了一种新的训练方法,将规则集成到深度学习中,在某种程度上,规则的强度在推理时是可控的。具有可控规则表示的深度神经网络(DeepCTRL)将规则编码器与基于规则的目标结合在一起,从而实现决策的共享表示。DeepCTRL对数据类型和模型体系结构是不可知的。它可以应用于为输入和输出定义的任何类型的规则。DeepCTRL的关键方面是它不需要重新训练来适应规则强度-在推理时,用户可以根据所需的操作点对准确度与规则验证比率进行调整。在现实世界的领域中,融入规则是至关重要的-例如物理,零售和医疗保健-我们展示了DeepCTRL在教授深度学习规则方面的有效性。DeepCTRL通过显着提高规则验证率来提高训练模型的信任度和可靠性,同时还在下游任务中提供准确性增益。此外,DeepCTRL还支持新的用例,例如对数据样本的规则进行假设测试,以及基于数据集之间共享规则的无监督自适应。

英文摘要:

We propose a novel training method that integrates rules into deep learning, in a way the strengths of the rules are controllable at inference. Deep Neural Networks with Controllable Rule Representations (DeepCTRL) incorporates a rule encoder into the model coupled with a rule-based objective, enabling a shared representation for decision making. DeepCTRL is agnostic to data type and model architecture. It can be applied to any kind of rule defined for inputs and outputs. The key aspect of DeepCTRL is that it does not require retraining to adapt the rule strength -- at inference, the user can adjust it based on the desired operation point on accuracy vs. rule verification ratio. In real-world domains where incorporating rules is critical -- such as Physics, Retail and Healthcare -- we show the effectiveness of DeepCTRL in teaching rules for deep learning. DeepCTRL improves the trust and reliability of the trained models by significantly increasing their rule verification ratio, while also providing accuracy gains at downstream tasks. Additionally, DeepCTRL enables novel use cases such as hypothesis testing of the rules on data samples, and unsupervised adaptation based on shared rules between datasets.

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。

↑