### Abstract

We present a framework for learning DFA from simple examples. We show that efficient PAC learning of DFA is possible if the class of distributions is restricted to simple distributions where a teacher might choose examples based on the knowledge of the target concept. This answers an interesting variant of an open research question posed in Pitt's seminal paper: Are DFA's PAC-identifiable if examples are drawn from the uniform distribution, or some other known simple distribution? Our approach uses the RPNI algorithm for learning DFA from labeled exampies. In particular, we describe an efficient learning algorithm for exact learning of the target DFA with high probability when a bound on the number of states (N) of the target DFA is known in advance. When N is not known, we show how this algorithm can be used for efficient PAC learning of DFAs.

Original language | English (US) |
---|---|

Title of host publication | Algorithmic Learning Theory - 8th International Workshop, ALT 1997, Proceedings |

Editors | Ming Li, Akira Maruoka |

Publisher | Springer Verlag |

Pages | 116-131 |

Number of pages | 16 |

ISBN (Print) | 3540635777, 9783540635772 |

DOIs | |

State | Published - Jan 1 1997 |

Event | 8th International Workshop on Algorithmic Learning Theory, ALT 1997 - Sendai, Japan Duration: Oct 6 1997 → Oct 8 1997 |

### Publication series

Name | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
---|---|

Volume | 1316 |

ISSN (Print) | 0302-9743 |

ISSN (Electronic) | 1611-3349 |

### Other

Other | 8th International Workshop on Algorithmic Learning Theory, ALT 1997 |
---|---|

Country | Japan |

City | Sendai |

Period | 10/6/97 → 10/8/97 |

### Fingerprint

### All Science Journal Classification (ASJC) codes

- Theoretical Computer Science
- Computer Science(all)

### Cite this

*Algorithmic Learning Theory - 8th International Workshop, ALT 1997, Proceedings*(pp. 116-131). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 1316). Springer Verlag. https://doi.org/10.1007/3-540-63577-7_39

}

*Algorithmic Learning Theory - 8th International Workshop, ALT 1997, Proceedings.*Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 1316, Springer Verlag, pp. 116-131, 8th International Workshop on Algorithmic Learning Theory, ALT 1997, Sendai, Japan, 10/6/97. https://doi.org/10.1007/3-540-63577-7_39

**Learning DFA from simple examples.** / Parekh, Rajesh; Honavar, Vasant.

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Learning DFA from simple examples

AU - Parekh, Rajesh

AU - Honavar, Vasant

PY - 1997/1/1

Y1 - 1997/1/1

N2 - We present a framework for learning DFA from simple examples. We show that efficient PAC learning of DFA is possible if the class of distributions is restricted to simple distributions where a teacher might choose examples based on the knowledge of the target concept. This answers an interesting variant of an open research question posed in Pitt's seminal paper: Are DFA's PAC-identifiable if examples are drawn from the uniform distribution, or some other known simple distribution? Our approach uses the RPNI algorithm for learning DFA from labeled exampies. In particular, we describe an efficient learning algorithm for exact learning of the target DFA with high probability when a bound on the number of states (N) of the target DFA is known in advance. When N is not known, we show how this algorithm can be used for efficient PAC learning of DFAs.

AB - We present a framework for learning DFA from simple examples. We show that efficient PAC learning of DFA is possible if the class of distributions is restricted to simple distributions where a teacher might choose examples based on the knowledge of the target concept. This answers an interesting variant of an open research question posed in Pitt's seminal paper: Are DFA's PAC-identifiable if examples are drawn from the uniform distribution, or some other known simple distribution? Our approach uses the RPNI algorithm for learning DFA from labeled exampies. In particular, we describe an efficient learning algorithm for exact learning of the target DFA with high probability when a bound on the number of states (N) of the target DFA is known in advance. When N is not known, we show how this algorithm can be used for efficient PAC learning of DFAs.

UR - http://www.scopus.com/inward/record.url?scp=84958043718&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84958043718&partnerID=8YFLogxK

U2 - 10.1007/3-540-63577-7_39

DO - 10.1007/3-540-63577-7_39

M3 - Conference contribution

AN - SCOPUS:84958043718

SN - 3540635777

SN - 9783540635772

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 116

EP - 131

BT - Algorithmic Learning Theory - 8th International Workshop, ALT 1997, Proceedings

A2 - Li, Ming

A2 - Maruoka, Akira

PB - Springer Verlag

ER -