#### DMCA

## Learning conjunctions of Horn clauses (1990)

### Cached

### Download Links

- [www.lirmm.fr]
- [www.lirmm.fr]
- [www-sal.cs.uiuc.edu]
- DBLP

### Other Repositories/Bibliography

Venue: | In Proceedings of the 31st Annual Symposium on Foundations of Computer Science |

Citations: | 121 - 13 self |

### Citations

1984 | A theory of the learnable - Valiant - 1984 |

771 | Learning quickly when irrelevant attributes abound: a new linear threshold algorithm
- Littlestone
- 1988
(Show Context)
Citation Context ...ins open, of whether general CNF and DNF formulas are learnable in polynomial time. Positive results in this line of research include improved algorithms for k-CNF and k-DNF formulas (Haussler, 1988; =-=Littlestone, 1988-=-), and algorithms for larger classes of formulas, including linearly separable formulas (Littlestone, 1988), internal disjunctive formulas (Haussler, 1988), decision lists (Rivest, 1987), and rank k d... |

755 |
Queries and concept learning.
- Angluin
- 1988
(Show Context)
Citation Context ...on, as PAC-learning general CNF and DNF formulas. One approach to finding efficient learning algorithms has been to relax the problem by making membership queries available to the learning algorithm (=-=Angluin, 1988-=-). That is, the learning algorithm is allowed to query domain elements of its choice and receive their correct classifications with respect to the target concept. Membership queries can make a differe... |

726 | Learnability and the Vapnik-Chervonenkis dimension - Blumer, Ehrenfeucht, et al. - 1989 |

427 | Learning decision lists
- Rivest
- 1987
(Show Context)
Citation Context ...er, 1988; Littlestone, 1988), and algorithms for larger classes of formulas, including linearly separable formulas (Littlestone, 1988), internal disjunctive formulas (Haussler, 1988), decision lists (=-=Rivest, 1987-=-), and rank k decision trees (Ehrenfeucht & Haussler, 1988). Kearns and Valiant (1989) give cryptographic evidence to tile effect that the class of all Boolean formulas is not polynomial-time learnabl... |

347 | Cryptographic limitations on learning Boolean formulae and finite automata. - Kearns, Valiant - 1994 |

253 |
Quantifying Inductive Bias: AI Learning Algorithms and Valiant's Model
- Haussler
- 1988
(Show Context)
Citation Context ...which still remains open, of whether general CNF and DNF formulas are learnable in polynomial time. Positive results in this line of research include improved algorithms for k-CNF and k-DNF formulas (=-=Haussler, 1988-=-; Littlestone, 1988), and algorithms for larger classes of formulas, including linearly separable formulas (Littlestone, 1988), internal disjunctive formulas (Haussler, 1988), decision lists (Rivest, ... |

181 | On the learnability of Boolean formulae - Kearns, Li, et al. - 1987 |

120 | When wonâ€™t membership queries help? - Angluin, Kharitonov - 1991 |

120 | Learning disjunctions of conjunctions - Valiant - 1985 |

115 | Learning read-once formulas with queries - Angluin, Hellerstein, et al. - 1989 |

107 | Negative results for equivalence queries. - Angluin - 1990 |

86 | Learning Conjunctive Concepts in Structural Domains. - Haussler - 1987 |

81 |
Complete Problems for Deterministic Polynomial Time.
- Jones, Laaser
- 1977
(Show Context)
Citation Context ...CNF, being "nearly monotone." For the class of propositional Horn sentences there are polynomial-time algorithms to decide the satisfiability of a formula and the logical equivalence of two formulas (=-=Jones & Laaser, 1977-=-), in contrast to the situation for general CNF formulas. A Horn clause such as (-1 a V --1 b V -~ c V d) is logically equivalent to an implication (a A b A c = d); thus, a propositional Horn sentence... |

74 |
Learning decision trees from random examples
- Ehrenfeucht, Haussler
- 1989
(Show Context)
Citation Context ...ms for larger classes of formulas, including linearly separable formulas (Littlestone, 1988), internal disjunctive formulas (Haussler, 1988), decision lists (Rivest, 1987), and rank k decision trees (=-=Ehrenfeucht & Haussler, 1988-=-). Kearns and Valiant (1989) give cryptographic evidence to tile effect that the class of all Boolean formulas is not polynomial-time learnable. Prediction-preserving reductions (Pitt & Warmuth, 1990)... |

60 | Predicting {0, 1} functions on randomly drawn points. - Haussler, Littlestone, et al. - 1994 |

49 | Predicting f0; 1g-functions on randomly drawn points - Haussler, Littlestone, et al. - 1990 |

42 |
Prediction-preserving reducibility.
- Pitt, Warmuth
- 1990
(Show Context)
Citation Context ...ucht & Haussler, 1988). Kearns and Valiant (1989) give cryptographic evidence to tile effect that the class of all Boolean formulas is not polynomial-time learnable. Prediction-preserving reductions (=-=Pitt & Warmuth, 1990-=-) provide a powerful method for relating the difficulties of distinct learning problems. For example, one of the reductions given by Kearns, Li, Pitt and Valiant (1987) shows that PAC-learning monoton... |

6 |
Learning with hints
- Angluin
- 1988
(Show Context)
Citation Context ...on, as PAC-learning general CNF and DNF formulas. One approach to finding efficient learning algorithms has been to relax the problem by making membership queries available to the learning algorithm (=-=Angluin, 1988-=-). That is, the learning algorithm is allowed to query domain elements of its choice and receive their correct classifications with respect to the target concept. Membership queries can make a differe... |

3 |
Requests for hints that return no hints
- Angluin
- 1988
(Show Context)
Citation Context ...on, as PAC-learning general CNF and DNF formulas. One approach to finding efficient learning algorithms has been to relax the problem by making membership queries available to the learning algorithm (=-=Angluin, 1988-=-). That is, the learning algorithm is allowed to query domain elements of its choice and receive their correct classifications with respect to the target concept. Membership queries can make a differe... |

1 | Learning read-once formulas with queries (Technical Report, University of California at - Angluin, Hellerstein, et al. - 1989 |