### Abstract

methods have recently been proposed for automatically learning accurate system models from data in the form of observations of the target system. Common for these approaches is that they assume the data to consist of multiple independent observation sequences. However, for certain types of systems, in particular many running embedded systems, one would only have access to a single long observation sequence, and in these situations existing automatic learning methods cannot be applied. In this paper, we adapt algorithms for learning variable order Markov chains from a single observation sequence of a target system, so that stationary system properties can be verified using the learned model. Experiments demonstrate that system properties (formulated as stationary probabilities of LTL formulas) can be reliably identified using the learned model.

Original language | English |
---|---|

Title of host publication | NASA Formal Methods : 4th International Symposium, NFM 2012, Norfolk, VA, USA, April 3-5, 2012. Proceedings |

Editors | Alwyn E. Goodloe, Suzette Person |

Number of pages | 15 |

Publisher | Springer |

Publication date | 2012 |

Pages | 216-230 |

ISBN (Print) | 978-3-642-28890-6 |

ISBN (Electronic) | 978-3-642-28891-3 |

DOIs | |

Publication status | Published - 2012 |

Event | NASA Formal Methods Symposium - Norfolk, United States Duration: 3 Apr 2011 → 5 Apr 2012 Conference number: 4 |

### Conference

Conference | NASA Formal Methods Symposium |
---|---|

Number | 4 |

Country | United States |

City | Norfolk |

Period | 03/04/2011 → 05/04/2012 |

Series | Lecture Notes in Computer Science |
---|---|

Volume | 7226 |

ISSN | 0302-9743 |

### Fingerprint

### Cite this

*NASA Formal Methods: 4th International Symposium, NFM 2012, Norfolk, VA, USA, April 3-5, 2012. Proceedings*(pp. 216-230). Springer. Lecture Notes in Computer Science, Vol.. 7226 https://doi.org/10.1007/978-3-642-28891-3_22

}

*NASA Formal Methods: 4th International Symposium, NFM 2012, Norfolk, VA, USA, April 3-5, 2012. Proceedings.*Springer, Lecture Notes in Computer Science, vol. 7226, pp. 216-230, NASA Formal Methods Symposium, Norfolk, United States, 03/04/2011. https://doi.org/10.1007/978-3-642-28891-3_22

**Learning Markov models for stationary system behaviors.** / Chen, Yingke; Mao, Hua; Jaeger, Manfred; Nielsen, Thomas Dyhre; Larsen, Kim Guldstrand; Nielsen, Brian.

Research output: Contribution to book/anthology/report/conference proceeding › Article in proceeding › Research › peer-review

TY - GEN

T1 - Learning Markov models for stationary system behaviors

AU - Chen, Yingke

AU - Mao, Hua

AU - Jaeger, Manfred

AU - Nielsen, Thomas Dyhre

AU - Larsen, Kim Guldstrand

AU - Nielsen, Brian

PY - 2012

Y1 - 2012

N2 - Establishing an accurate model for formal verification of an existing hardware or software system is often a manual process that is both time consuming and resource demanding. In order to ease the model construction phase, methods have recently been proposed for automatically learning accurate system models from data in the form of observations of the target system. Common for these approaches is that they assume the data to consist of multiple independent observation sequences. However, for certain types of systems, in particular many running embedded systems, one would only have access to a single long observation sequence, and in these situations existing automatic learning methods cannot be applied. In this paper, we adapt algorithms for learning variable order Markov chains from a single observation sequence of a target system, so that stationary system properties can be verified using the learned model. Experiments demonstrate that system properties (formulated as stationary probabilities of LTL formulas) can be reliably identified using the learned model.

AB - Establishing an accurate model for formal verification of an existing hardware or software system is often a manual process that is both time consuming and resource demanding. In order to ease the model construction phase, methods have recently been proposed for automatically learning accurate system models from data in the form of observations of the target system. Common for these approaches is that they assume the data to consist of multiple independent observation sequences. However, for certain types of systems, in particular many running embedded systems, one would only have access to a single long observation sequence, and in these situations existing automatic learning methods cannot be applied. In this paper, we adapt algorithms for learning variable order Markov chains from a single observation sequence of a target system, so that stationary system properties can be verified using the learned model. Experiments demonstrate that system properties (formulated as stationary probabilities of LTL formulas) can be reliably identified using the learned model.

UR - http://www.scopus.com/inward/record.url?scp=84859453785&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-28891-3_22

DO - 10.1007/978-3-642-28891-3_22

M3 - Article in proceeding

SN - 978-3-642-28890-6

SP - 216

EP - 230

BT - NASA Formal Methods

A2 - Goodloe, Alwyn E.

A2 - Person, Suzette

PB - Springer

ER -