paper-with-me

홈 › Papers

Stability of Internal States in Recurrent Neural Networks Trained on Regular Languages

2020-06-18 · Christian Oliva, Luis F. Lago-Fernández

We provide an empirical study of the stability of recurrent neural networks trained to recognize regular languages. When a small amount of noise is introduced into the activation function, the neurons in the recurrent layer tend to saturate in order to compensate the variability. In this saturated regime, analysis of the network activation shows a set of clusters that resemble discrete states in a finite state machine. We show that transitions between these states in response to input symbols are deterministic and stable. The networks display a stable behavior for arbitrarily long strings, and when random perturbations are applied to any of the states, they are able to recover and their evolution converges to the original clusters. This observation reinforces the interpretation of the networks as finite automata, with neurons or groups of neurons coding specific and meaningful input patterns.

📄 PDF Abstract BibTeX arXiv:2006.10828

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Representing Formal Languages: A Comparison Between Finite Automata and Recurrent Neural Networks

2019-02-27 · Joshua J. Michalenko, Ameesh Shah, Abhinav Verma, Richard G. Baraniuk 외

We investigate the internal representations that a recurrent neural network (RNN) uses while learning to recognize a regular formal language. Specifically, we train a RNN on positive and negative examples from a regular …

Clustering

Noisy Recurrent Neural Networks

2021-02-09 · NeurIPS 2021 12 · Soon Hoe Lim, N. Benjamin Erichson, Liam Hodgkinson, Michael W. Mahoney

We provide a general framework for studying recurrent neural networks (RNNs) trained by injecting noise into hidden states. Specifically, we consider RNNs that can be viewed as discretizations of stochastic differential …

General Classification

Finite Automata Can be Linearly Decoded from Language-Recognizing RNNs

2019-05-01 · ICLR 2019 5 · Joshua J. Michalenko, Ameesh Shah, Abhinav Verma, Swarat Chaudhuri 외

We study the internal representations that a recurrent neural network (RNN) uses while learning to recognize a regular formal language. Specifically, we train an RNN on positive and negative examples from a regular langu…

Clustering

A Dynamically Controlled Recurrent Neural Network for Modeling Dynamical Systems

2019-10-31 · Yiwei Fu, Samer Saab Jr, Asok Ray, Michael Hauser

This work proposes a novel neural network architecture, called the Dynamically Controlled Recurrent Neural Network (DCRNN), specifically designed to model dynamical systems that are governed by ordinary differential equa…

State Space Models

Asymptotic Stability in Reservoir Computing

2022-06-07 · Jonathan Dong, Erik Börve, Mushegh Rafayelyan, Michael Unser

Reservoir Computing is a class of Recurrent Neural Networks with internal weights fixed at random. Stability relates to the sensitivity of the network state to perturbations. It is an important property in Reservoir Comp…