1) What is the difference between
MultiWordNet and EuroWordNet?
Like MultiWordNet, EuroWordNet is a project aimed at the development of wordnets for the main European languages. EuroWordNet was an EU-funded project which involved different partners, among which was the Istituto di Linguistica Computazionale of Pisa where the Italian component was developed. The EuroWordNet project is not related to MultiWordNet, which is being independently developed at ITC-irst.
The two multilingual lexica basically differ in the procedure followed in the construction of the language specific wordnets. In the EuroWordNet project, the language specific wordnets have been built independently from each other, trying in a second phase to find correspondences between them. On the contrary, in MultiWordNet the language specific wordnets have been built keeping as much as possible of the semantic relations available in the Princeton WordNet. This is done by building the new synsets in correspondence with the PWN synsets whenever possible, and importing semantic relations from the corresponding English synsets; i.e., we assume that if there are two synsets in PWN and a relation holding between them, then the same relation holds between the corresponding synsets in the new language.
|
 |
2) Is the MultiWordNet database accessed or downloaded on-line?
Yes, MultiWordNet can be freely accessed on-line from the web site you are visiting, though the files are distributed for licensed use only.
|
|
3) Are there different kinds of licences available and how much do they cost?
All information about the different kinds of licences and their costs are available on the web page "Licence".
|
|
4) Which languages constitute the MultiWordNet database?
The on-line version includes seven aligned languages: English (Princeton WordNet), Italian (developed by the TCC group at ITC-irst, now FBK), Spanish (developed by the TALP group at Universitat Politenica de Catalunya), Portuguese (developed by the NLX-Group at the University of Lisbon), Hebrew (developed by the University of Haifa in collaboration with ITC-irst, now FBK), Romanian (developed by the "Alexandru Ioan Cuza" University of Iasi) and Latin (developed by the University of Verona).
The MultiWordNet distributed version contains, just, the Italian database aligned with the English Princeton WordNet. The Spanish database can be directly acquired from the TALP group at Universitat Politenica de Catalunya.
|
|
5) Can only companies and institutions or also private users obtain a licence ?
Private users can also get a licence.
|
|
6) What do I have to do and how long does it take to me to obtain a licence?
The data of the purchasing institution or company and its representative in the licence are required, i.e. first and last names, qualification, institution/company's address, telephone, fax, e-mail. Once we get the data, our office will send you a proposal of licence by e-mail and if you agree to it, we will send you the original to be signed and sent back to us. As soon as we receive the signed original licence, MultiWordNet is made available for use and a payment invoice is sent to you. Usually, the whole procedure does not take more than one month.
|
|
7) Is the MultiWordNet licence transferable to other partners within the same project?
No, the MultiWordNet licence holds only for the institution which signs it and is not transferable under any condition.
|
|
8) Do you distribute an XML version of MultiWordNet?
There is an XML version, but we usually distribute the dump of our database in SQL format: text files which allow for the construction of a database to interface with (we use MySQL). Further information about the distribution data format are available on the web page "Distribution data format".
|
|
9) Do you distribute an RDF version of MultiWordNet?
No implementation of an RDF version of MultiWordNet is planned for the moment.
|
|
10) Is it possible to interface MultiWordNet with Java?
Yes, the Java API are available on request.
|
|
11) Are MultiWordNet new releases freely distributed to the owners of a licence?
Every new version of MultiWordNet will be signalled in the NEWS section of the website and in the MultiWordNet Forum. The latest version of MultiWordNet will be made available to the owners of a licence on request.
|
|
12) Is a stand alone browsing interface available?
Not for the moment, but in the future it will be available.
|
|
13) What do you think about a Wiki Wiki Web approach to MultiWordNet?
In our opinion, the Wiki Wiki Web approach does not guarantee the reliability of the resource. MultiWordNet data are manually checked by expert lexicographers highly trained at working on this resource. Often the consequences of a change in the synonyms or in a relation are not understood by a person who does not have an overall view of the organization of WordNet.
What is really of primary importance, instead, is to receive bug reports, comments, and advices of any kind so that further improvements can be made within the resource. To contribute to the improvement of MultiWordNet please become a registered user (see the "Registration" page).
|
|
14) What is the difference between MultiWordNet and WordNet-Domains?
WordNet-Domains is a data base in which each synset of the English WordNet has been annotated with at least one domain label (e.g. Medicine, Sport, Politics). Information brought by domains is complementary to what is already in WordNet. This extension of the original WordNet provide a useful way to establish semantic relations among word senses, which can be profitable for many uses, in particular during the automatic disambiguation process. A domain may include synsets from different syntactic categories: for instance Medicine groups together senses from Nouns, such as doctor#1 and emergency_room#1, and from Verbs such as operate#7. WordNet-Domains is composed of a text file containing the id number of all the synsets in English WordNet, and for each of them, the lexical category together with the assigned domain label are reported: for example "00081298-n baseball". This resource can be freely obtained through the compilation of the licence available from the WordNet-Domains web site.
The MultiWordNet lexical database distributed by FBK is composed of the Italian database in which each synset (set of synonyms representing a lexical concept) is aligned with the corresponding synset in the English Princeton WordNet. Words and synsets are then linked by various lexical and semantic relations (hypo/hypernymy, meronymy, antonymy, etc.). As domains are language-independent, domain labels are present also in the Italian WordNet and thus they are included in MultiWordNet too.
|
|
: mwnstaff
fbk.eu
