DeepAI AI Chat
Log In Sign Up

Weighted Automata Extraction from Recurrent Neural Networks via Regression on State Spaces

by   Takamasa Okudono, et al.

We present a method to extract a weighted finite automaton (WFA) from a recurrent neural network (RNN). Our algorithm is based on the WFA learning algorithm by Balle and Mohri, which is in turn an extension of Angluin's classic algorithm. Our technical novelty is in the use of regression methods for the so-called equivalence queries, thus exploiting the internal state space of an RNN. This way we achieve a quantitative extension of the recent work by Weiss, Goldberg and Yahav that extracts DFAs. Experiments demonstrate that our algorithm's practicality.


page 1

page 2

page 3

page 4


Extracting Finite Automata from RNNs Using State Merging

One way to interpret the behavior of a blackbox recurrent neural network...

Learning Deterministic Weighted Automata with Queries and Counterexamples

We present an algorithm for extraction of a probabilistic deterministic ...

Discovering Novel Customer Features with Recurrent Neural Networks for Personality Based Financial Services

The micro-segmentation of customers in the finance sector is a non-trivi...

Connecting Weighted Automata and Recurrent Neural Networks through Spectral Learning

In this paper, we unravel a fundamental connection between weighted fini...

Connecting Weighted Automata, Tensor Networks and Recurrent Neural Networks through Spectral Learning

In this paper, we present connections between three models used in diffe...

Distillation of Weighted Automata from Recurrent Neural Networks using a Spectral Approach

This paper is an attempt to bridge the gap between deep learning and gra...

Analyzing Recurrent Neural Network by Probabilistic Abstraction

Neural network is becoming the dominant approach for solving many real-w...