Top  UNL Expression   Relations   Attributes  Universal Words  UNLKB  Knowledge Representation in UNL  Logical Expression in UNL  UNL System

UNL 2005 Specifications

7 June 2005
Copyright © UNL Center of UNDL Foundation

Chapter 8  UNL System

The UNL System consists of three major components: language resources, software for processing the language resources, and tools and systems for maintaining and operating the language processing software or language resources. Language resources are divided into language dependent part and language independent part. Linguistic knowledge on concepts that universal to every language is considered language independent and to be stored in the common database UNLKB. Language dependent resources like word dictionaries and analysis and generation rules, as well as the software for language processing, are stored in each language server. Language servers are connected through the Internet. Supporting tools for producing UNL documents can be used in a local PC. Such supporting tools operate with consulting language servers through the Internet. Verification of UNL documents can be carried out through the Internet or in a local PC. UW Gate for searching and maintaining the common database UNLKB operates through the Internet.

8.1  Structure of the UNL System

In figure 8.1, highlighted parts show the components of the UNL System, white parts show applications and their UNL database based on the UNL System.

Figure 8.1 Structure of the UNL System

Each component of figure 8.1 is the following:

UNLKCIC stores information of Key Concept in Context (KCIC) about UNL documents. The UNLKCIC is made for every binary relation of UNL documents. This information is used for searching related UNL expressions of a UNL expression. Through UNLKCC, every UW of UNLKB is linked to the UNL documents each UW involved, and all UWs included in the UNL documents are also stored in the UNLKB and linked to corresponding UNL documents. UNLKCIC expands the knowledge in UNLKB from merely linguistic knowledge such as binary relations between context independent concepts to real world knowledge such as binary relations between context sensitive concepts restricted by other concepts knowledge.

Concept Definitions are the collection of UNL documents made for the sentences of definitions of UWs. These definitions of concepts provide the knowledge of concepts in connection with other concepts that can specify the concepts. This knowledge is indispensable for reasoning in information retrieval, etc.

UNL Documents mean the documents in which UNL expression is described for each sentence of natural language. A UNL document can be made of a plain text file or an UNL-embedded html file. A UNL document base is a collection of UNL document files. UNL documents are for the purpose to provide information and knowledge in UNL expression.

UNLKB is a semantic network comprising every directed binary relation between UWs. These binary relations are the possible relations that each UW can have with others. Such relations are established based on the UW System. Such UNLKB has the following functions: 1) defines semantics (concepts) of UWs, and 2) provides linguistic knowledge of concepts. For details see Chapter 4 Universal Words and Chapter 5 UNLKB.

UW Dictionary stores and provides the interface between UWs and words of natural languages.

UNL Verifier verifies whether a UNL expression is correct syntactically, lexically and semantically. The syntax check of a UNL expression is carried out against the UNL Specifications. In lexical check, whether all UWs of a UNL expression are defined in the UNLKB are checked. In semantic check, whether each binary relation of a UNL expression is defined as possible is certified with consulting the UNLKB.

UNL Language Servers (LSs) are located in the Internet to carry out the conversions between natural languages and UNL expressions. Each LS contains an EnConverter and a DeConverter of a language. EnConverter converts natural language sentences to UNL expressions. DeConverter convert UNL expressions to natural language sentences.

UW Gate is a tool for people to access the UNLKB and the UW dictionary through the Internet.

UNL Proxy Server works in a local computer to communicate with language servers. It functions as a filter to check whether a web page that a user required is written in UNL or not. If UNL expressions are included in the web page, it communicates with an appropriate language server in the Internet for deconverting the UNL expressions into desired language sentences and provides the Internet browser with the results to display.

UNL Editor is a tool helping to produce UNL documents. It includes an EnConverter and a DeConverter. Each of them can be selected according to language. EnConverter converts natural language sentence into UNL expressions. DeConverter provides generated results as feedback for checking the correctness of UNL expressions.

UNL Explorer provides the basic means to knowledge infrastructure. It manages UNLKB, UNLKCIC and UNL documents and provides knowledge or information through UWs. The UNL Explorer can be used in two ways. For human, it allows users or developers to view or to develop the UNL Knowledge System such as the UNL Encyclopedia. For computers, it provides information or knowledge on UWs.

The UNL Explorer uses UNLKB for navigating information stored in UNL database. It has two windows: the hierarchy of UWs (UW System) of UNLKB is shown in the left window. UWs of the UNLKB are keys for information stored in UNL database. Information on UWs is shown in the right window through navigation through the UW System. Information on UWs is described in UNL documents. All UWs used in the UNL documents are included in the left window of UNLKB and are keys for further information.

The UNL Explorer allows users to search for information using UWs or words of natural languages. It shows the information in UNL or a desired natural language by accessing UNL Language Servers. It also provides functions for developers to add information to or modify information of the UNL database in their native languages.

Information about a UW is stored in a file. Location of the file is linked with the UW. This architecture of the UNL Knowledge System allows its development to be carried out by a wide range of developers from different languages and cultures. Such a database can provide a wealth of up-to-date information on various aspects of information and knowledge from all over the world.

UNL Encyclopedia is a collection of UNL documents describing various knowledge or information. The merit of the UNL Encyclopedia is that it enables each people to use its native language to share knowledge or information with other peoples in their respective languages, through UNL Explorer for example. The UNL Encyclopedia basically describes ordinary encyclopedias in UNL and integrates those descriptions (UNL documents) into the UNLKB by linking the UNL documents to corresponding UWs. It is a multilingual encyclopedia navigating through the UNLKB for human, with the knowledge description necessary for computers to process knowledge.

UNL Library is an UNL-based application providing bibliographic information (e.g. MARC21) and contents (in future) in UNL, and allowing users to search for books through UNL.

8.2  Mechanism of conversion of UNL expressions

Figure 8.2 shows the mechanism how a UNL document is made and how a UNL document is converted into natural languages in the UNL system. Arrows in solid line show dataflow, arrows in broken line show access.

The EnConverter and DeConverter are the core software in the UNL system. The EnConverter converts natural language sentences into UNL expressions. The Universal Parser (UP) is a specialized version of the EnConverter. It generates UNL expressions from annotated sentences using the UW dictionary without using grammatical features. All UNL expressions are verified by the UNL verifier. The DeConverter converts UNL expressions to natural language sentences.

Figure 8.2 Mechanism of conversion of UNL

Top  UNL Expression   Relations   Attributes  Universal Words  UNLKB  Knowledge Representation in UNL  Logical Expression in UNL  UNL System