Extracting automata from neural networks using active learning.
Extracting automata from neural networks using active learning.
复制标题
使用主动学习从神经网络中提取自动机
DOI：
10.7717/peerj-cs.436
复制
发表时间：
2021
期刊：
影响因子：
--
通讯作者：
He M
中科院分区：
文献类型：
--
作者：
Xu Z;Wen C;Qin S;He M
深度学习是机器学习最先进的形式之一。大多数现代深度学习模型都基于人工神经网络，基准研究表明，神经网络产生的结果与人类专家相当，在某些情况下甚至上级人类专家。然而，生成的神经网络通常被认为是不可理解的黑箱模型，这不仅限制了其应用，也阻碍了测试和验证。在本文中，我们提出了一个主动学习框架，提取自动机的神经网络分类器，这可以帮助用户理解的分类器。更详细地说，我们使用Angluin的L* 算法作为学习者，学习下的神经网络作为一个预言机，采用抽象解释的神经网络回答成员资格和等价查询。我们的抽象包括价值、符号和词语的抽象。文中还讨论了影响提取的因素。我们已经在原型中实现了我们的方法。为了对其进行评估，我们在MNIST分类器上执行了原型，并确定间隔数为2且块大小为1 × 28的抽象在F1得分方面提供了最佳性能。我们还将我们提取的DFA与通过LearnLib中提供的被动学习算法学习的DFA进行了比较，实验结果表明，我们的DFA在MNIST数据集上具有更好的性能。
Deep learning is one of the most advanced forms of machine learning. Most modern deep learning models are based on an artificial neural network, and benchmarking studies reveal that neural networks have produced results comparable to and in some cases superior to human experts. However, the generated neural networks are typically regarded as incomprehensible black-box models, which not only limits their applications, but also hinders testing and verifying. In this paper, we present an active learning framework to extract automata from neural network classifiers, which can help users to understand the classifiers. In more detail, we use Angluin’s L* algorithm as a learner and the neural network under learning as an oracle, employing abstraction interpretation of the neural network for answering membership and equivalence queries. Our abstraction consists of value, symbol and word abstractions. The factors that may affect the abstraction are also discussed in the paper. We have implemented our approach in a prototype. To evaluate it, we have performed the prototype on a MNIST classifier and have identified that the abstraction with interval number 2 and block size 1 × 28 offers the best performance in terms of F1 score. We also have compared our extracted DFA against the DFAs learned via the passive learning algorithms provided in LearnLib and the experimental results show that our DFA gives a better performance on the MNIST dataset.
登录
查看更多内容
影响因子：
7.4
作者：
FUJIWARA, S;BOCHMANN, GV;GHEDAMSI, A
通讯作者：
GHEDAMSI, A
影响因子：
1
作者：
ANGLUIN, D
通讯作者：
ANGLUIN, D
影响因子：
2.9
作者：
TINO, P;SAJDA, J
通讯作者：
SAJDA, J
影响因子：
2.9
作者：
ZENG, Z;GOODMAN, RM;SMYTH, P
通讯作者：
SMYTH, P
影响因子：
7.8
作者：
Omlin, CW;Giles, CL
通讯作者：
Giles, CL