OLIMPO SYSTEM WEB-TECNOLOGY FOR ELECTR

ONIC

GOVERNMENT AND WORLD PEACE

Hugo C. Hoeschl, Tânia Cristina D. Bueno, Vânia Barcellos, Andre Bortolon, Eduardo S.

Mattos

E-Gov, Juridical Intelligence and Systems Institute – Ijuris

Keywords: UN Security Council´s Resolutions, Dynamically Contextualised Knowledge Representation (DCKR),

Structured Contextual Search – SCS, Information of Technology, Data retrieve, JAVA

Abstract: The paper describes the Olimpo System, a knowledge-based system that enables the user to access textual

files and to retrieve information that is similar to the search context described by the user in natural

language. The paper is focused on the innovation recently implemented on the system and its new features.

A detailed description is presented about the search level and the similarity metrics used by the system. The

methodology applied to the Olimpo system emphasises the use of information retrieval methods combined

with the Artificial Intelligence technique named SCS (Structured Contextual Search).

1 INTRODUCTION

Some complex and specific domains require an

information retrieval system that is more than just a

great technology to search for documents in large

text databases. A good knowledge representation is

also required.

The present approach enables to retrieve textual

information that is similar to the search text

described by the user using natural language.

Through the extraction of relevant information using

DCKR technology (Dynamically Contextualised

Knowledge Representation) [8] [9], new documents

are automatically included in the knowledge

database. Concepts of Case-Based Reasoning (CBR)

[1] [2] and information retrieval techniques were

applied to obtain a better performance of the system,

leading to the technology named Structured

Contextual Search – SCS.

The following item 2 of this paper addresses the

UN Security Council and the Resolution document;

in items 3 and 4 the knowledge representation

methodology is presented and Olimpo system is

described; in items 5 and 6, describe how the web-

technology will be applied to Olimpo system and the

impacts that influence not only in the technological

field, but also in relation to the citizenship, the

knowledge and the research and item 7 is the

conclusion of the paper.

2 ABOUT THE UN SECURITY

COUNCIL

The importance of the UN body becomes noticeable

when one follows the main global means of

communication and no further arguments are

required. Being the source of the documents handled

by the Olimpo system, it is useful to give more

details about the Security Council and its document

base.

According to its Charter (Article 7-1), the United

Nations Organization (UNO) is comprised of six

special bodies. All of them issue relevant

documentation and it is highly important to have an

adequate tool to retrieve those documents.

Given its characteristics and aspects related to

the Resolutions, the Security Council was chosen as

application field of the Olimpo system.

The Security Council is described by Article 7-1

of UNO’s Charter, where it is referred to as a special

body of the United Nations. The Security Council is

specifically addressed in Chapter V, Articles 23 to

32. As per Article 24-1, its central function is to

assume “the main responsibility in maintaining

international peace and security.”

It should be emphasized that the Security

Council has a juridical and an executive profile.

According to Kelsen (apud Steinfus [11]), it is

juridical because it holds the monopoly of legitimate

289

C. Hoeschl H., Cristina D. Bueno T., Barcellos V., Bortolon A. and S. Mattos E. (2004).

OLIMPO SYSTEM WEB-TECNOLOGY FOR ELECTRONIC GOVERNMENT AND WORLD PEACE.

In Proceedings of the Sixth International Conference on Enterprise Information Systems, pages 289-296

DOI: 10.5220/0002644902890296

 SciTePress

violence at the international scope and judges the

existence of facts, determines sanctions on them and

who will enforce these sanctions. That turns it a

juridical body. And this profile enables a good

application of the technology of juridical

information, especially SCS and its particular

method of rhetoric structure analysis of a given

jurisdictional context, based on the knowledge

structure involving the body, which maximizes the

task of intelligent retrieval of documents when

adequate modelling is used.

The Security Council has also political

characteristics and it has discritionary power to

establish violations, according to Steinfus [11];

therefore the Security Council holds an executive

characteristic, turning it a juridical-executive body.

The Security Council presents some peculiarities.

One of them is to be currently the most powerful

jurisdictional body on the planet. Another one is the

existence of internal, informal instances, named “P

3” (Western permanent member countries) and “P 5”

(all permanent member countries), according to

Steinfus [11]. Another peculiarity is the existence of

internal bodies with specific power delegation to

perform certain tasks, on a permanent or “ad hoc”

level, like the sanctions committee, as shown on

Figure 1.

General Assembly

Committees

Security Council

Permanent

“Ad hoc”

Sanctions

“P 3”

“P 5”

Plenary

• Principle of “big power

unanimity”, or

•“Veto power”

“Instances”

Source: www.un.org

Figure 1: Position of the Security Council

Among the documents issued by the Security

Council, six of them have greater relevance. Based

on their structure and relevance, the Resolutions

were chosen for the application of the Olimpo

system.

As per the structure of the document, the

Resolutions have some characteristics that make it

easier to apply the technology referred to herein

3 DYNAMICALLY

CONTEXTUALISED

KNOWLEDGE

REPRESENTATION

Olimpo´s performance is centred about the

combination of aspects derived from CBR and text

information retrieval, in addition to an adequate

organization of the knowledge related to the subject

the system is focused on (in the present case, the UN

Security Council´s Resolutions). The

aforementioned knowledge organization is what

enables the DCKR technology, which is a

methodology that provides the possibility of

comparing the contexts described in the documents

and not only a comparison between words or

attributes.

3.1 Analysis of the rhetoric structure

The rhetoric structure of the system is comprised of

indicative expressions used for comparison means

and it was, first time ever, dynamically prepared. Up

to then it was usual to choose a list of index pointers

from a source external to the research group (for

example, Court library indexes). Little work was

done on the list of index pointers and its selection

was based on its similarity with the context of the

system under development. For the Olimpo system it

was decided to build a particular and specific list,

which should be aligned with the issues effectively

ICEIS 2004 - ARTIFICIAL INTELLIGENCE AND DECISION SUPPORT SYSTEMS

290

treated by the Resolutions and, on the other hand,

should be coherent with the documentation context

of the managing entity of the database. In this view,

in order to collect a list of expressions a detailed

reading of the Resolutions was performed, searching

onto UNO´s database on the Internet was done and

debating with research groups was used. Those

expressions were then tested and subject to statistic

analysis in order to evaluate their function as

reference elements for the indexation and retrieval of

documents. A set of expressions with high

significance was selected, eliminating those ones

with very high or very low frequency of occurrence

because they were not very much helpful for

establishing a context.

This process had a dynamic characteristic

because it was done several times and expressions

were included or excluded according to their statistic

performance. The routine described shows how it

worked to obtain a final list containing a set of

expressions that could efficiently reflect the generic,

rhetoric structure of the Resolutions, which gave the

material form to the dynamically contextualised

knowledge representation.

3.2 Structured contextual search –

SCS

The searching process being described is said to be

“contextual” and “structured” due to the following

reasons:

a) For building the rhetoric structure of the system,

it is taken into consideration the context of the

stored documents;

b) This context is the basis for the input adjustment

process, as well as for the comparison and

selection of documents;

c) When writing the search text, the input is not

limited to a set of words or attributes, but it can

take the format of a long text, including the

possibility of setting specific attributes, which

work as filters and function as a preliminary

selection of documents to be searched.

Information contained in the documents is

represented in the form of a case, consisting of the

original document and a set of eight indexes in the

form of pairs of attribute-value: subject, date,

number of the Resolution, meeting, country,

acronyms, decisions, and indicative expressions.

These indexes are part of the system interface (see

Figure2).

Figure 2: Olimpo Interface

In general, the system works in a way similar to

other case-based systems [3] [5] (see Figure 3),

where a manual entry passes through an adjustment

and is then submitted to a comparison with the

documents contained in the database, from which

the most suitable ones are selected based on

similarity calculations

OLIMPO SYSTEM WEB-TECNOLOGY FOR ELECTRONIC GOVERNMENT AND WORLD PEACE

291

Figure 3: Searching Process

After a refined modelling of the database, the

Resolutions are stored by Olimpo system, according

to their characteristics and central attributes (main

topics, related subjects, countries involved);

peripheral attributes (other related Resolutions, other

UNO´s organisms referred to); and superficial

attributes (dates, numbers and names). This kind of

structure allows to give (variable) weights to

attributes, enabling a more precise, contextualised

search.

Furthermore, the control of depth of search

enables a selection of documents according to a

higher or lower occurrence of indicative expressions

within the text of the Resolution, before starting to

compare the documents. This process provides a

more efficient way of reducing the search field; it is

not a mere pre-selection of documents based on their

superficial characteristics, but a preliminary

comparison oriented by the context related to the

search input.

After completing the process, the result is a list

of indicative expressions referring to the

Resolutions, producing an individual record of the

occurrence of each one of the expressions within the

text of each Resolution. These records allow the

system to make the comparison and to apply the

global similarity metrics.

In addition to the indicative expressions, the

process of automatic extraction of attributes was

prepared to detect and extract the subject, date,

number of the Resolution, acronyms, country names,

and parts of the text that contain the expressions

with higher occurrence.

The main features of the Olimpo system are the

simultaneous use of textual information retrieving

techniques based on CBR and the possibility of an

extensive textual input. That makes the system to

reach a differentiated performance in terms of

information retrieval.

However, considering that the comparisons are

based on a rhetoric structure previously prepared,

the better working of the system is linked to a

description of the search entry closer to that rhetoric

structure. So, the system performance becomes

gradually more consistent as the search entry

language gets closer to the structure identified in the

documents that generated the knowledge base of the

system.

It has to be mentioned that all the Resolutions

were monitored all the time with respect to the

number of indicative expressions they presented

during the structuring phase of the Resolutions

knowledge base.

4 SIMILARITY METRICS

The similarity metrics was structured to consider the

indicative expressions present in the case and in the

search, after applying the rhetoric structure on the

textual entry and producing the derived search. This

derived search is actually the reference to work out

the similarity metrics.

Taking as an example the case where a derived

search with a total of 50 expressions is obtained after

applying the rhetoric structure on a given search

text: this set of expressions is compared to the

records in the database and the similarity percentage

is calculated based on the number of similar

expressions found within each individual record. If

43 expressions are found, for instance, then the

2) Applying the rhetoric structure onto the manual entry

3) Obtaining a derived search

5) Comparison is performed and Resolutions

are selected

1) Manual entry

4) The derived search is submitted

to the database

6) Output processing

7) Most similar Resolutions are shown

ICEIS 2004 - ARTIFICIAL INTELLIGENCE AND DECISION SUPPORT SYSTEMS

292

similarity will be 86%; it will be 72%, if 36

expressions are found, or 56% in the case of finding

just 28 expressions, and so on.

This type of metrics is quite simple, one of the

most simple that could be used in this situation, but

it works in a quite stable way and can be improved

in the future by incorporating new mechanisms like

trigrams or applying internal weights to the most

frequent words found in the text of the Resolution.

In fact, what is the strong feature of the system is

not the similarity metrics, but the way how the

indicative expressions are organized so that the

metrics provides a better performance.

A clear example of this particularity of the

system is an expression formed by two words like

“United Nations”. A simple similarity based on

counting individual words will show a 100% index

when both words are found within the text,

regardless their position, or 50% in the case just one

of the words is found. However, if a differentiated

indexation is used, by which “United Nations” (the

two exact words appearing together) is one

expression, “United” is another expression, and

“Nations” is a third one (all with the same weight,

for the time being), this configures a different

situation. In this case, it is not enough finding the

two words within the text, even when separated;

they should appear together and having the exact

meaning. Based on these criteria, the similarity

index will be 33.33% when only one of the two

words is found, or 66.66%, when both words are

found in separate location, and it will reach 100%

only when both words are present and appear

together.

5 THE FUTURE OF OLIMPO

SYSTEM

Internet was developed more than three decades ago,

financed by the Department of the Defense of the

U.S.A. Originally projected to connect the main

systems of computers about a dozen of universities

and organizations of research, the internet is

currently accessible in hundreds of millions of

computers in the world.

With the introduction of the World Wide Web -

which allows personal computers to visualize

documents based on multimedia about almost all the

subjects - the internet literally explodes for what it

certainly seems to become, the main mechanism of

communication of the world.

Most computer applications were executed in

computers which couldn´t communicate between

themselves. Currently, applications that are

communicated with hundreds of millions of

computers of the world can be written. The internet

establishes technology of computation and

communications.

It turns our work easier. The information is

accessible world-wide in an instantaneous and

convenient form. It becomes possible that

individuals and local small companies have a world-

wide exposition. The people can find the better

prices in products or services. Communities that

have similar interests can stay in contact with each

other. Researchers can be informed instantaneously

about de last advances around the globe.

JAVA applications can be written in any

computer platform, it means, any Java application

needs a 32 bits version, such as Windows 95,

Windows NT, other premium versions of Microsoft

operational systems, MAC and UNIX, what results

in a important economy of time and costs of

development of systems for enterprises.

Java is a language completely object-oriented

with hard support for proper techniques of

engineering softwares. The programming object-

oriented shapes objects of the real world with

corresponding softwares. It takes advantage of the

class relationships in which objects of a certain type

- as a type of vehicles - they have the same

characteristics. It also takes advantage of inheritance

relationships where just-brought objects of a certain

type inheritate characteristics of existing ranks, but

still keeping exclusive characteristics. An object of

the convertible rank certainly has the characteristics

of the automobile class, but the ceiling of a

convertible opens and closes.

The object-oriented programme (OOP) supplies

us a more natural and intuitive way to see the

process of programming – as follows, shaping

objects of the real world, its attributes and its

behaviors. The OOP also shapes the communication

between objects. In the same ways as people change

messages between themselves, the objects also

communicate by messages.

The OOP encapsulates data (attributes) and

methods (behaviour) in packages called objects, data

and methods of an object are intimately united. The

objects have the property of hiding the information.

This means that even though the objects can

communicate with each other through friendly

interfaces, normally it is not allowed for objects to

know how the other objects are implemented - the

implementation details are occulted inside of the

objects themselves. Certainly it is possible to direct a

car without knowing well the details of how the

engine, the transmission and the system of exhaust

pipe work internally.

For the use on the WEB, the interfaces of

Olimpo® system will be adapted in JAVA, because

of the advantages shown above. The data base will

OLIMPO SYSTEM WEB-TECNOLOGY FOR ELECTRONIC GOVERNMENT AND WORLD PEACE

293

be extended, so as to enclose all the Resolutions of

the Security Council, making it possible for the users

of the Web to carry through consultations to the

complete base and thus to take off more advantage

of the system. The site will be of free access to the

public, or either, without no cadastre form, and

moreover it will possess a system of automatic

updating to facilitate maintenance of information.

The system will have basically two distinct

modules: the first one will be of the consultation

that, based in the methodology of Context Structured

Search – CSS will allow a similar retrieval to the

existing one in the current system, containing filters

and fields to open-search concept-based, and the

administration module, who will be responsible for

the inclusion of new resolutions and maintenance of

the database. The administration module is

necessary, because Olimpo® system has a base of

knowledge especially developed to retrieve in an

intelligent way the resolutions of Security Council.

Knowledge base is a referential structure of

representation in that domain specially studied, built

so that the algorithms to retrieve information, can be

semantic references in the search.

The knowledge base of Olimpo® system will be

opened after its implementation to insert news data.

Those can be given through the insertion of new

resolutions, as well as more words and references in

the specialized dictionary, which is meant for the

mining and retrieval of information contained in the

system database.

The updating of the data will be made in a

dynamic form, that is, all the process will be made

with a reduced effort, being up to the administrator

of the system the function of including of new data

in the base and the managing of the dynamics of the

working of the system, keeping this up to date and

compatible with the new referring events the

Security Council of the ONU.

As follows below, there is a sketch of how the

structure of the net will be . (Figure 4):

Figure 4

Server

MRE's Workstation

BRASILIA - MRE

(Ministério das

Relações Exteriores)

MRE's foreigner offices

Database

UN Security

Council´s

Resolutions

New York

ICEIS 2004 - ARTIFICIAL INTELLIGENCE AND DECISION SUPPORT SYSTEMS

294

6 FORESEEN IMPACTS

Application of Olimpo® system will in such a way

bring a series of impacts at national and international

level, influencing not only in the technological field,

but also in relation to citizenship, the knowledge and

the research.

See some relevant impacts bellow:

Scientific Impact

• Accomplishment of seminaries and workshops

on the applicability of the Technology of the

Legal Information for the Security Council of the

ONU;

• Publications in newspapers of the Institutions

involved;

• Stimulus to the perfectioning of the techniques

applied;

• Academic works about this theme.

Technological Impact

• Consolidation of the technique “Dynamically

Contextualised Knowledge Representation”

(RCKR), which increases performance of

systems structured in knowledge;

• Perfectioning of the techniques of use of textual

data bases with RCKD, that propitiate the

application of PCE, for construction of Olimpo®

system;

• Projection of Brazil as a technological reference.

Economic Impact

• Promotion of the companies of technology of the

Information and Management of Knowledge;

• Dissemination of the applicability of the

technologies used in Olimpo® by areas, that

present the same difficulty in document research;

• Dissemination of the techniques of applied

Management of the Knowledge to Olimpo®.

Social Impact

Facing the efficiency and the agility presented

with the use of Olimpo® in the processes of search

of information, increasing the search of the users to

the available services.

7 CONCLUSION

Olimpo system is a clear example of an innovative

approach to the issue of information retrieval from

complex text databases. Based on CSS technology

(Context Structured Search, the system reaches a

higher performance using DCKR technique for

knowledge representation.

The innovation and new features implemented

represent an upgrade of Olimpo system, improving

its overall performance and usability.

Olimpo® System will be an important Brazilian

contribution for the unanimity between nations, in

view of the fact that it democratizes the access to

knowledge of the select circle of the Security

Council of the ONU, what makes it an important

technological contribution for the World-wide

Peace. Faster, necessary and perfect decisions will

have, surely, greater and better international

legitimation, favoring the creation of a fairer scene.

As a consequence, it will be an instrument for the

demonstration of the technological capacity withheld

by Brazil, in the area of softwares for the

management of knowledge.

REFERENCES

Amondt, A.; Plaza, E. “Case-Based Reasoning:

Fundamental Issues, Methodological Variations, and

System Approaches”. AI Communications 17(1),

1994.

Bench-Capon, T. J. M. Some observations on modelling

case based reasoning with formal argument models.

In: Proceedings of the Seventh International

Conference on Artificial Intelligence and Law, p. 36-

42, Oslo: Norway, June 14-18, 1999. 220 p.

Bruninghaus, Stefanie; ASHLEY, Kevin D. Toward

adding knowledge to learning algorithms for indexing

legal cases. In: Proceedings of the Seventh

International Conference on Artificial Intelligence and

Law, p. 9-17, Oslo: Norway, June 14-18, 1999. 220 p.

Bueno, Tânia Cristina D´Agostini. The use of juridical

theory for retrieval from large juridical textual

databases. Master Dissertation, PPGEP/UFSC.

Florianópolis (Brazil): 1999. Original title: O uso da

teoria jurídica para recuperação em amplas bases de

textos jurídicos.

Bueno, Tania Cristina D'Agostini; Hoeschl, Hugo Cesar;

Mattos, Eduardo da Silva; Barcia, Ricardo Miranda;

Wangenheim, Christiane Gresse Von. JurisConsulto:

Retrieval in Jurisprudencial Text Bases using Juridical

Terminology. In: The Seventh International

Conference on Artificial Intelligence And Law, 1999,

Oslo. Proceedings of the Conference. New York:

ACM, 1999. v.1. p.147-155.

Bueno, Tania Cristina D'Agostini; Hoeschl, Hugo Cesar;

Mattos, Eduardo da Silva; Wangenheim, Christiane

Gresse Von; Barcia, Ricardo Miranda. The use of

juridical theory for retrieval from large juridical

textual databases. In: Encontro Nacional de

Inteligência Artificial, 1999, Rio de Janeiro. Anais do

XIX Congresso Nacional da Sociedade Brasileira de

OLIMPO SYSTEM WEB-TECNOLOGY FOR ELECTRONIC GOVERNMENT AND WORLD PEACE

295

Computação. Rio de Janeiro: Edições EntreLugar,

1999. v.4. p.107-120. Original title: Uso da teoria

jurídica para recuperação em amplas bases de textos

jurídicos.

Bueno, Tania Cristina D'Agostini; Hoeschl, Hugo Cesar;

Mattos, Eduardo da Silva; Barcia, Ricardo Miranda;

Bortolon, André; Wangenheim, Christiane Gresse

Von. JurisConsulto. Florianópolis (Brazil): 1999.

Software rights registered.

Hoeschl, Hugo Cesar. Olimpo System: Juridical

Information Technology for UNO´s Security Council.

Florianópolis (Brazil): UFSC, 2002. Doctorate Thesis.

Original title: Sistema Olimpo: Tecnologia da

Informação Jurídica para o Conselho de Segurança da

ONU.

Hoeschl, Hugo Cesar; Barcia, Ricardo Miranda; Bueno,

Tânia Cristina D´Agostini; Mattos, Eduardo da Silva;

Bortolon, Andre; Donatti, Fabrício Tadeu. Olimpo

System. Florianópolis (Brazil), 2000. Software rights

registered.

Hoeschl, Hugo Cesar; BUENO, Tania Cristina D'agostini;

MATTOS, Eduardo da Silva; BORTOLON, André;

RIBEIRO, Marcelo Stopanowski; THEISS, Irineu;

BARCIA, Ricardo Miranda. STRUCTURED

CONTEXTUAL SEARCH FOR THE UN

SECURITY COUNCIL. In: ICEIS - 5TH

INTERNATIONAL CONFERENCE ON

ENTERPRISE INFORMATION SYSTEMS, 2003,

Angers. Proceedings of the fifth International

Conference On Enterprise Information Systems.

Setúbal: School of Technology of Setúbal, 2003. v. 2,

p. 100-107. Referências adicionais: Classificação do

evento: Internacional; França/Inglês; Meio de

divulgação: Impresso; Homepage:

http://www.iceis.org/papers.htm; ISSN/ISBN:

9729881618.

Hoeschl, Hugo Cesar; BUENO, Tania Cristina D'agostini;

BORTOLON, André; RIBEIRO, Marcelo

Stopanowski; MATTOS, Eduardo da Silva; THEISS,

Irineu. Dynamically contextualized knowledge

representation of the United Nations Security Council

Resolutions. In: NINTH INTERNATIONAL

CONFERENCE ON ARTIFICIAL INTELLIGENCE

AND LAW, 2003, Edimburgo. ICAIL 2003

Proceedings. New York: ACM, 2003. v. 1, p. 95-96.

Referências adicionais: Classificação do evento:

Internacional; Escócia/Inglês; Meio de divulgação:

Impresso; Homepage:

http://www.cirfid.unibo.it/~agsw/icail03/; ISSN/ISBN:

1581137478.

Hoeschl, Hugo Cesar; Bueno, Tânia Cristina D´Agostini;

Mattos, Eduardo da Silva; Bortolon, Andre; Barcia,

Ricardo Miranda. Olimpo: Contextual Structured

Search to improve the representation of UN Security

Council Resolutions with information extraction

methods. In: The 8th International Conference on

Artificial Intelligence and Law, 2001, St.Louis, MO,

USA. Proceedings of the Conference. New York:

ACM, 2001. v.1. p. 271-218.

Steinfus, Ricardo. Handbook of International

Organizations. Porto Alegre (Brazil), 1997, 352p.

Original title: Manual de organizações internacionais.

ICEIS 2004 - ARTIFICIAL INTELLIGENCE AND DECISION SUPPORT SYSTEMS

296