## A model of information retrieval based on a terminological logic (1993)

### Cached

### Download Links

- [nmis.isti.cnr.it]
- [www.nmis.isti.cnr.it]
- [faure.iei.pi.cnr.it]
- [nmis.isti.cnr.it]
- [faure.iei.pi.cnr.it]
- [faure.isti.cnr.it]
- [gaia.isti.cnr.it]
- DBLP

### Other Repositories/Bibliography

Citations: | 94 - 19 self |

### BibTeX

@INPROCEEDINGS{Meghini93amodel,

author = {Carlo Meghini and Fabrizio Sebastiani and Umberto Straccia and Costantino Thanos},

title = {A model of information retrieval based on a terminological logic},

booktitle = {},

year = {1993},

pages = {298--307},

publisher = {ACM Press}

}

### Years of Citing Articles

### OpenURL

### Abstract

According to the logical model of Information Retrieval (IR), the task of IR can be described as the extraction, from a given document base, of those documents d that, given a query q, make the formula d → q valid, where d and q are formulae of the chosen logic and “→ ” denotes the brand of logical implication formalized by the logic in question. In this paper, although essentially subscribing to this view, we propose that the logic to be chosen for this endeavour be a Terminological Logic (TL): accordingly, the IR task becomes that of singling out those documents d such that d � q, where d and q are terms of the chosen TL and “�” denotes subsumption between terms. We call this the terminological model of IR. TLs are particularly suitable for modelling IR; in fact, they can be employed: 1) in representing documents under a variety of aspects (e.g. structural, layout, semantic content); 2) in representing queries; 3) in representing lexical, “thesaural ” knowledge. The fact that a single logical language can be used for all these representational endeavours ensures that all these sources of knowledge will participate in the retrieval process in a uniform and principled way. In this paper we introduce Mirtl, a TL for modelling IR according to the above guidelines; its syntax, formal semantics and inferential algorithm are described. 1

### Citations

906 |
A Framework for Representing Knowledge
- Minsky
- 1975
(Show Context)
Citation Context ...of early knowledge representation languages (of which KL-ONE [3] may be considered the founding father) based on semantic networks and inspired by the notion of frame, originally introduced by Minsky =-=[14]-=- in the context of cognitive science applications. Quite apart from AI domains, these early languages have also been applied in conceptual data modelling (see e.g. [10; 18]) and conceptual document mo... |

281 | An analysis of first-order logics of probability
- Halpern
- 1990
(Show Context)
Citation Context ...gs can be addressed by remaining within the paradigm of TLs. To this respect, we are currently considering the idea of introducing probabilistic reasoning in our model along the approach described in =-=[8]-=-. The computational complexity of the reasoning algorithm is also a major problem that has to be tackled 12 . From this point of view, we are already treading on dangerous ground, as Mirtl is a proper... |

240 | R.: Semantic database modelling: Survey, applications and research issues
- Hull, King
- 1987
(Show Context)
Citation Context ...originally introduced by Minsky [14] in the context of cognitive science applications. Quite apart from AI domains, these early languages have also been applied in conceptual data modelling (see e.g. =-=[10; 18]-=-) and conceptual document modelling [11]. Unfortunately these languages did not possess a formal semantics; hence, the fact that the notion of inference they enforced were the same as the user expecte... |

235 | The complexity of concept languages
- Donini, Lenzerini, et al.
- 1991
(Show Context)
Citation Context ...he literature (see e.g. [20]). For our purposes, it will suffice to say that some of these results are encouraging, in the sense that for some (reasonable) choice of Σ, optimal logics have been found =-=[6]-=- that have also a fair amount of expressive power 8 . 3 The Terminological Model of IR. An informal introduction. Up to now, we have given an informal account of what a TL is and what are the main pro... |

183 |
A non-classical logic for Information Retrieval
- Rijsbergen
- 1986
(Show Context)
Citation Context ...L for modelling IR according to the above guidelines; its syntax, formal semantics and inferential algorithm are described. 1 Introduction According to the logical model of Information Retrieval (IR) =-=[21; 22]-=-, the task of IR can be described as the extraction, from a given document base, of those documents d that, given a query q, make the formula d → q valid, where d and q are formulae of the chosen logi... |

172 |
Entailment, the logic of relevance and necessity
- Anderson, Belnap
- 1975
(Show Context)
Citation Context ...uestion.sALEN logic 13 , which is known to be NP-hard [7]. In order to solve this problem we are currently working [12] on a variant of Mirtl based, rather than on classical logic, on relevance logic =-=[1]-=-; early results in the field suggest that “relevance” TLs are in general computationally easier to handle than their classical equivalents [17]. An alternative solution to the complexity problem that ... |

161 | Semantic Data Models
- Peckham, Maryanski
- 1988
(Show Context)
Citation Context ...originally introduced by Minsky [14] in the context of cognitive science applications. Quite apart from AI domains, these early languages have also been applied in conceptual data modelling (see e.g. =-=[10; 18]-=-) and conceptual document modelling [11]. Unfortunately these languages did not possess a formal semantics; hence, the fact that the notion of inference they enforced were the same as the user expecte... |

149 |
Attributive concept descriptions with complements. Artiflcial Intelligence 48:1{26
- Schmidt-Schau…, Smolka
- 1991
(Show Context)
Citation Context ...r all individuals denoted by the latter are also denoted by the former. For example, such a theorem prover will be able to answer a rmatively to the question whether the term 4 Some authors (see e.g. =-=[7, 19]-=-) call \terminological" only those TLs that do not allow instance assertions� these authors call \hybrid logics" those TLs that do allow such assertions. 300 (and polygon (atmost 4 sides)) (which deno... |

143 |
Attributive concept description with complements
- Smolka, Schmidt-Schauss
(Show Context)
Citation Context ...)) [Fido] dog [Fido] 3 For an informal explanation of the meaning of the termforming operators involved in the examples that follow, the reader may peek ahead in Section 4.1. 4 Some authors (see e.g. =-=[7; 19]-=-) call “terminological” only those TLs that do not allow instance assertions; these authors call “hybrid logics” those TLs that do allow such assertions. author [John,paper2501] We have remarked befor... |

107 |
Terminological cycles: semantics and computational properties
- Nebel
- 1990
(Show Context)
Citation Context ... terminological axioms are acyclic, i.e. that their left-hand side does not occur, either directly or indirectly, in their right-hand side. This is a standard simplifying assumption for TLs; see e.g. =-=[15]-=- for a framework that does away with this limitation.s1. a terminological module (TM), consisting of the operators for forming complex terms; 2. a definitional module (DM), consisting of the “ . =” an... |

68 |
A Structural Paradigm for Representing Knowledge
- Brachman
- 1977
(Show Context)
Citation Context ...telligence research and, in particular, in the subfield of AI known as “knowledge representation”; TLs derive, in fact, from a large class of early knowledge representation languages (of which KL-ONE =-=[3]-=- may be considered the founding father) based on semantic networks and inspired by the notion of frame, originally introduced by Minsky [14] in the context of cognitive science applications. Quite apa... |

56 |
A new Theoretical Framework for Information Retrieval
- Rijsbergen
- 1986
(Show Context)
Citation Context ...L for modelling IR according to the above guidelines; its syntax, formal semantics and inferential algorithm are described. 1 Introduction According to the logical model of Information Retrieval (IR) =-=[21; 22]-=-, the task of IR can be described as the extraction, from a given document base, of those documents d that, given a query q, make the formula d → q valid, where d and q are formulae of the chosen logi... |

47 |
An information retrieval model based on modal logic
- Nie
- 1989
(Show Context)
Citation Context ...ns should be much more expressive and richly structured than the language of simple statements adopted by (either the classical or the non-classical variants of) propositional logic and considered in =-=[16; 21; 22]-=-. This latter language is also inadequate in another important respect. The basic linguistic unit of propositional logic is the statement, an expression whose denotation in a given interpretation is a... |

41 | About retrieval model and logic - Chiaramella, Chevallet - 1992 |

36 |
Conceptual modeling of multimedia documents
- Meghini, Rabitti, et al.
- 1991
(Show Context)
Citation Context ...matical logic, be at the same time adequate for accounting for the complexity of real documents and queries, especially when multimedia documents are considered. In fact, it has been argued (see e.g. =-=[11]-=-) that a radical improvement in the performance of IR systems may only be accomplished by endowing these systems with high level representations of the documents contained in the document base. In ord... |

31 | Readings in Knowledge Representation - Brachman, Levesque - 1985 |

22 | A four-valued semantics for frame-based description languages
- Patel-Schneider
- 1986
(Show Context)
Citation Context ...based, rather than on classical logic, on relevance logic [1]; early results in the field suggest that “relevance” TLs are in general computationally easier to handle than their classical equivalents =-=[17]-=-. An alternative solution to the complexity problem that we are considering is the investigation of probabilistic algorithms for subsumption checking. The answer that these algorithms would give to th... |

22 | Attributive concept descriptions with complements - Schmidt-Schaub, Smolka - 1991 |

16 |
A formal definition for the expressive power of knowledge representation languages
- Baader
- 1990
(Show Context)
Citation Context ...o use a TL in an IR modelling context, yielding the terminological model of IR. 6 For an attempt at giving a more general and rigorous definition of “expressive power of a TM”, see the work by Baader =-=[2]-=-. 7 For the sake of simplicity, we will assume that none among the operators in Σ may be expressed by means of a combination of other operators of Σ; this will allow us to say that different subsets a... |

16 |
Y.: An analysis of rst-order logics of probability
- Halpern
- 1990
(Show Context)
Citation Context ...gs can be addressed by remaining within the paradigm of TLs. To this respect, we are currently considering the idea of introducing probabilistic reasoning in our model along the approach described in =-=[8]-=-. The computational complexity of the reasoning algorithm is also a major problem that has to be tackled 12 . From this point of view, we are already treading on dangerous ground, as Mirtl is a proper... |

10 |
From subsumption to instance checking
- Donini, Lenzerini, et al.
- 1992
(Show Context)
Citation Context ...)) [Fido] dog [Fido] 3 For an informal explanation of the meaning of the termforming operators involved in the examples that follow, the reader may peek ahead in Section 4.1. 4 Some authors (see e.g. =-=[7; 19]-=-) call “terminological” only those TLs that do not allow instance assertions; these authors call “hybrid logics” those TLs that do allow such assertions. author [John,paper2501] We have remarked befor... |

6 | A computationally tractable terminological logic
- Sebastiani, Straccia
- 1990
(Show Context)
Citation Context ...ofitably used in computationally demanding applications. A number of results relative to the computational complexity of subsumption checking in TMs have recently appeared in the literature (see e.g. =-=[20]-=-). For our purposes, it will suffice to say that some of these results are encouraging, in the sense that for some (reasonable) choice of Σ, optimal logics have been found [6] that have also a fair am... |

2 |
A formal de nition of the expressive power of knowledge representation languages
- Baader
- 1990
(Show Context)
Citation Context ...tion for TLs� see e.g. [15] for a framework that does away with this limitation. 6 For an attempt at giving a more general and rigorous de nition of \expressive power of a TM", see the work by Baader =-=[2]-=-.sde ned as the \ " partial order de ned on the powerset of 7 . It is well known that there is a tradeo between the expressive power of a logic and its computational tractability� hence, in general, t... |

2 | bertoStraccia.A computationallytractableterminologicallogic - Um |

1 |
Towards a relevance logic of information retrieval. Unpublished manuscript
- Meghini, Sebastiani
- 1992
(Show Context)
Citation Context ...l use it as a first approximation of the computational feasibility of the logics in question.sALEN logic 13 , which is known to be NP-hard [7]. In order to solve this problem we are currently working =-=[12]-=- on a variant of Mirtl based, rather than on classical logic, on relevance logic [1]; early results in the field suggest that “relevance” TLs are in general computationally easier to handle than their... |

1 | Donini,MaurizioLenzerini,Daniele Nardi,andW ernerNutt.The complexity ofconcept onealwaysrefersto\worstcasecomplexity";althoughinan IR contextthismightnotbean interestingparameter,inthe absenceofstatisticalcharacterizationsofthe\averagecase" we w - FrancescoM - 1991 |

1 | Haugeland,editor.Mind design. The MIT Press,Cam bridge,MA - John - 1981 |

1 | Conceptualmodelingofm ultimediadocuments - CarloMeghini, Thanos |

1 | InPatrick J.Winston,editor,The psychology ofcomputervision - frameworkforrepresentingknowledge - 1975 |

1 | Attributive conceptdescriptionswithcomplements.Artificial Intelligence,48:1-26,1991 - GertSmolka |

1 | Entailment - the logic of reieuance and necessity - Anderson, Belnap - 1975 |

1 | A new theoretical framework for information retrieval - Rljsbergen - 1986 |