GENERAL KNOWLEDGE MACHINE PROJECT

ELECTRONIC KNOWLEDGE PUBLISHING

ELECTRONIC KNOWLEDGE SYSTEM AUTHOR'S GUIDE

by Konstantin M Golubev
revised 20-Feb-2013

It is strongly recommended to take a look at General Knowledge Machine Research Group Web site : http://gkm-ekp.sf.net.

Index

Overcoming limitations
General Knowledge Machine
E-knowledge system development with General Knowledge Machine

Development of Windows version
Licensing

Overcoming limitations

Let us think about the following. We are all humans, and therefore limited in our intellectual abilities. Existing and emerging knowledge is great in volume, even in very specific areas like medicine. No one can learn it all, or even greater part. But knowledge is working only when it is in our minds. In reality it means that in many cases opportunities are not applied because we did not learn needed knowledge. Though every one of us would like to get support based on the best knowledge available, e.g. get right diagnosis and treatment. Our research group proposes to consider two new approaches for this.


1. Electronic knowledge publishing

We treat it as a further development of traditional and electronic publishing. Electronic knowledge systems are based on human knowledge found in external sources and really multiply human intellectual abilities with facilities for effective immediate consulting. They might be considered as a sound replacement of artificial intelligence expert systems and neural networks, since they fit readily into our intellectual activity and are built on totally human-oriented approach. Please read the details in a paper "Is there any future for artificial intelligence?" published in 2001 issue of ISPIM News (International Society for Professional Innovation Management, http://www.ispim.org. You may find electronic version at http://gkm-ekp.sf.net/gkm-ai.htm. Also you may read a paper "Introducing a new kind of publishing: The Electronic Knowledge Publishing" at http://gkm-ekp.sf.net/gkmekp.htm.

2. Adaptive learning

It is based on a concept called Just-In-Time Knowledge. It means that e-knowledge systems may serve as intelligent forefront to any knowledge source including living people, providing effective search of knowledge suitable exactly for a given situation. They provide this way ability to learn only knowledge really needed. Please read the details in a paper "Adaptive learning for knowledge-based organization" published in February 2002 by SSGRR (Scuola Superiore G. Reiss Romoli) international e-business conference in Italy (http://www.ssgrr.it/en/ssgrr2002w/papers.htm). Electronic version may be found at http://gkm-ekp.sf.net/gkm-ssgrr.htm

Practical implementation

Both approaches were used to build application systems in medicine, management consulting, banking, arts and Web sites. Seven papers are published on e-knowledge systems theory and implementation in Russia, Japan, Italy, Great Britain and France.

We propose collaboration on mutual projects development in the following areas:

  1. E-knowledge systems creation in different areas of scientific and business life. These systems could be used for immediate consulting and adaptive learning.
  2. Methodology of e-knowledge systems creation. Development of e-knowledge systems methodology, manuals, courses.
  3. Creation of international community on e-knowledge systems for mutual projects development and experience exchange.

Please contact us by email : gkm-ekp@users.sf.net
or by snail mail : Mailbox 33, Kiev-191, 03191, Ukraine


Warm regards, Konstantin M Golubev, Maria Gavrilina

General Knowledge Machine Research Group http://gkm-ekp.sf.net

List of published papers

1. E.T.Mikhailenko, V.I.Milko, A.N.Kostyuchenko, K.M.Golubev, V.A.Sokolova. Computers in the X-Ray Diagnosis. Midwifery and gynaecology, 1989, No.10, Moscow, Russia

2. N.B.Mankovsky, I.N.Karaban,G.N.Kryzhanovsky, V.A.Evseev, S.V.Magaeva, I.A.Vetrile, N.A.Trekova, L.A.Basharova, K.M.Golubev. Dopamine antibodies in Parkinsonian patients and their possible role in the Parkinson's syndrome pathogenesis. Neurology and Psychiatry Magazine named after S.S.Korsakov, v.93., 1993, No. 6, Moscow, Russia.

3. G.N.Kryzhanovsky, N.B.Mankovsky, I.N.Karaban, S.V.Magaeva, N.A.Trekova, I.A.Vetrile, L.A.Basharova, M.A.Atadzhanov, K.M.Golubev. Serotonin antibodies and their possible role in Parkinsonism. Neurology and Psychiatry Magazine named after S.S.Korsakov, v.94., 1994, No. 5, Moscow, Russia.

4. K.M.Golubev. Traditional + Adaptive Learning = broad way to Knowledge. Distance learning advancement project. Proceedings of ISPIM'99 conference in Tokyo, Japan, 1999.

5. K.M.Golubev. Is there any future for Artificial Intelligence?. ISPIM News, Ferbruary 2001

6. K.M.Golubev. Adaptive learning with e-knowledge systems. Scheduled for publication in the International Journal for Technology Management. Inderscience. 2002

7. K.M.Golubev. Adaptive learning for knowledge-based organization. Published as official paper of SSGRR-2002 Conference in Italy

General Knowledge Machine

Aim: Research, development and introduction of advanced approach to knowledge presentation and distribution called Electronic Knowledge Publishing. It is intended to provide transformation of individual knowledge into knowledge publicly accessible and usable for consulting and adaptive learning.

General Knowledge Machine is a set of tools, dedicated to effective knowledge search, on-line knowledge-based consulting and adaptive learning. It is possible to embed it in any application needed intelligent support including Web sites. There are additional development tools for user interface development, but at this time they are not free. Please contact author by email : gkmgeosite@yahoo.com.

Comparison of AI Expert Systems and E-knowledge Systems.

Expert Systems. E-knowledge Systems.
AI - Intended to replace human experts.EKP - Intended to assist human intellect.
AI - Based primarily on mathematics.EKP - Based on neurophysiology, psychology, knowledge management theory and mathematics.
AI - It is practically impossible to transform directly external knowledge sources to expert systems.EKP - It is further advancement of a traditional publishing - external knowledge sources (books, articles etc) may be transformed into e-knowledge systems easily.
AI - Based on the decision rules concept.EKP - Based on the general knowledge concept.
AI - The more complex an expert system is - the worse it works.EKP - The more complex e-knowledge system is - the better it works.
AI - A development has many stages and very expensive.EKP - A development has one stage and relatively inexpensive.
AI - It is relatively hard work to incorporate an expert system into other information systems due to sequential nature of data input and output.EKP - E-knowledge system may be easily incorporated into any kind of information system due to support of wide range of data input and output sources.
AI - It is practically impossible to use expert systems for learning, because they are not based on human knowledge.EKP - It may be easily used as a forefront to Adaptive Distance Learning applications, based on the Just-In-Time Knowledge concept.
AI - May not be used for new knowledge creation.EKP - May be used for new knowledge creation.
AI + EKP - Good news for all AI developers: the most valuable resource - knowledge bases on the productions concept, may be used for electronic knowledge systems development.AI+EKP


Hardware and software requirements and limitations

General Knowledge Machine e-knowledge base engine ('soz.exe' and 'gkm.exe') is written in GNU Fortran compiler. You may compile it for all hardware and software supported by GNU compiler project (tested in Windows, Linux and Unix environments). This engine is developed under GNU Lesser GPL license and therefore may be used in any projects including commercial. Sources and binaries may be found at gkm-ekp.sourceforge.net

Software compilation

In order to compile for your system e-knowledge base engine provide that GNU compiler is on your system (visit http://www.gnu.org).

Download and unpack to selected directory sources. Change directory to selected. Issue commands :

g77 soz.f -o soz.exe
g77 gkm.f -o gkm.exe

E-knowledge creation

Application 'soz.exe' should be used to create e-knowledge base (set of .gkm files). Start it with command 'soz.exe' in Windows, or './soz.exe' in Linux/Unix. Application 'gkm.exe' is interface between software applications and e-knowledge base. You may find information about interface in chapter 'Development of Windows version'. Initial sources have limitations : up to 2,000 possible signs, up to 8,000 possible situations (etalons), up to 100 signs in problem description, up to 100 signs in situation (etalon) description, up to 1,000 size of e-knowledge base element. It is sufficient to build very large e-knowledge system. But if it is necessary to build greater system, you should edit file 'param.prm' and recompile sources.

You may find e-knowledge sources for 'Language of gestures' project in packages of sources and binaries (files 'gestures.*').

E-knowledge system development with General Knowledge Machine.

ELECTRONIC KNOWLEDGE SYSTEM (E-knowledge system) is a knowledge-based system which helps human intellect to be more powerful and creative. It was developed to assist during process of intellectual activity, providing knowledge-based consulting and adaptive learning tools.

Step I

E-knowledge system is able to accept a knowledge accumulated by people, and this is a basis of it's power. Information about knowledge should be prepared for input into e-knowledge system in a special way. You should name the project usually using up to 8 symbols suitable for file name (for example, e-knowledge system on language of gestures may be named 'gestures'). Create a directory where system will reside, for example, 'd:\gestures'. Copy there all development tools (soz.exe and gkm.exe).

Preliminarily, it is needed to remove from initial speech or text all excessive information, keeping titles, exact descriptions of situations and recommendations how to handle them. Please try to concentrate on ideas rather than on words. We propose to define 'idea' as a standard text unequivocally and directly defining a specific side of a situation i.e. representing a stable structure in a brain's right part responsible for working with images of the world. This text should not include any excessive words and the words included should always have the right sense. You should control appearance of synonymous ideas, carefully removing duplications that are usually great in number. Resulting text, which may be used as example, in a case of language of gestures information from the book "Body language" by A.V.Pease, is the following.

**************** START OF TEXT *******************************
1.Consent.
----------
Complete description

Nodding.

2.Disagreement.
---------------
Complete description

Shaking one's head from side to side.

...

17.Self-defense pose, uncertainty, nervousness.
-----------------------------------------------
Complete description

Crossed hands on a breast.
The person holds himself by a hand.
Adjusting of cuff links, watches, buttons.
Pressing to oneself a handbag, a purse.
Wineglass is hold by two hands.
Putting one leg on another leg.
Fixation of one leg's foot on shank of another one.

Recommendations

Give a handle, book, paper, etc. into hands.
Ask to bend ahead, to consider any thing.
...

and so on.
**************** END OF TEXT *********************************

Every idea in a description of situation we will call a sign. In principle, any sign can have a grade, for example, if a sign is 'Weight' it may have grades '1-Very low, 2-Low, 3-Mean, 4-High, 5-Very high.'. We strongly recommend use 5 grades wherever it's possible, if number of grades is greater than 1. But in our example all signs have only one grade. From our experience, very rarely there's need to use sign with several grades.

Step II

You should gather all signs from all descriptions of situations, eliminate duplicates, and number them, pointing number of possible grades. We should make a file called, for example, 'gestures.sgn' of the following type.

**************** START OF FILE *******************************
1.Nodding. - It's comments - (Sign's title up to 64 symbols )
1 ,(Number of sign's grades)
2.Shaking one's head from side to side.
1,
3.Smile.
1,
4.Opened palms.
1,
5.Raised shoulders.
1,

...

and so on.
**************** END OF FILE *********************************

Step III

You should make a file called for example 'gestures.stn', containing numbered titles of possible situations accompanying with numbers of corresponding signs and grades, where zero means end of list. If sign's grade is equal to 1 than you may just place one more comma. It should look like the following.

**************** START OF FILE *******************************
1.Consent. - It's comments - (Title up to 64 symbols)
1,1, (Sign's number, grade's number)
0 (End of list)
2.Disagreement.
2,2,
0

...

20.Attempt to conceal deception when speaking.
21,,22,,42,,43,,44,,47,,55,,
0

...

and so on.
**************** END OF FILE *********************************

Step IV

You should make a file containing a questionnaire for a convenient description of problem. In this questionnaire you should group questions any way regardless of its numbers. It is allowed to include additional explanations, if it's needed. Please note that the first line for Internet version should be the name of Web server directory there the system will reside (for this example '/gestures'), format of chapter's title is '== Title', format of sign's number is 'NNNN~' (for example ' 11~) and format of sign's grade is 'NN~'. You should make a file called, for this example, 'gestures.que' of the following type.

**************** START OF FILE *******************************
/gestures
== Electronic Knowledge System: Language of gestures
GKM Research Group /gkmgeosite@yahoo.com/
Konstantin M Golubev
Ideas : Pease A.V., Body language, 1981, Australia, ISBN N 095936580 OX
----------------------------------------------------------------
Psychologists state that during a dialogue up to 40% of the information is
transmitted by words and an intonation, and more than 60% by gestures,
therefore it is very important to understand the language of gestures and
movements, which are not supervised by consciousness and are exact
indicators of state of a person.
ELECTRONIC KNOWLEDGE SYSTEM is intended to help you
- To define that your partner tells lie;
- To transform partners into the adherents;
- To distinguish expression of internal negative state and to smooth it;
- To inspire partners with trust and sympathy;
- To keep leadership in negotiations and dialogue.
-----------------------------------------------------------------------
Choose appropriate attributes, and you will see the list of the
propositions reflecting state of your partners.
== EYES =============================================================
--- EYES' PUPILS ---
77~Extended eyes' pupils.
78~Narrowed eyes' pupils.
--- EYEBROWS, EYELIDS ---
6~Raised eyebrows.
85~Slightly raised eyebrows.
86~eyebrows lowered downwards.
89~Eyelids closely cover eyes.
--- LOOKING ---
44~Taking one's eyes off.
79~Partner's eyes meet with yours less than 1/3 of dialogue time.
80~Partner's eyes meet with yours more than 2/3 of dialogue time.
81~Look in a forehead and eyes of partner.

...

and so on.
**************** END OF FILE *********************************

Step V

You should make a file containing all propositions regarding identified situations. You should call it gestures.prp' and it should look like the following. The first line of proposition is a title of situation from file 'gestures.stn' preceding by '~' sign. After this proposition text is going. Please note that for Internet version of e-knowledge system you may include any HTML tags into proposition text (references to pictures, multimedia, URL and so on).

**************** START OF FILE *********************************
~ 1.Consent.
-------------

Complete description

<EM>1.Nodding.</EM>
~ 2.Disagreement.
------------------

Complete description

2.Shaking one's head from side to side.
...

~ 93.Competing defensive position while negotiating.
----------------------------------------------------

Complete description

117.Arrangement of the participants of negotiations opposite over a table.

Recommendations

Proceed to following positions:
An arrangement of the participants of negotiations beside
each other over a corner of a table.
Arrangement of the participants of negotiations beside at a table.

...

and so on.
**************** END OF FILE *********************************

Step VI

You should make a file called 'soz.ini' of the following format:
Title of e-knowledge system
Signs descriptions file name (.sgn)
Situations descriptions file name (.stn)
Questionnaire file name (.que)
3,
0,
0,
For example, in our system:

**************** START OF FILE *********************************
Electronic Knowledge System: Language of gestures
gestures.sgn
gestures.stn
gestures.que
3,
0,
0,

**************** END OF FILE *********************************

Please run command interpreter. Change directory to appointed for development. All previously created files must be placed in this directory. After that run the program 'soz.exe'. On completion view the file 'proto.soz'. If it does not contain errors messages than initial e-knowledge base creation (files *.gkm) was successful. If there are errors please edit your files.

You should make a file called gkm.ini' of the following format:
Title of e-knowledge system
db.gkm
db2.gkm
Questionnaire file name (.QUE)
3,
Maximum grades deviation (%),
Minimum value index (%),
0,0,0,

For this example:

**************** START OF FILE *********************************
Electronic Knowledge System: Language of gestures
db.gkm
db2.gkm
gestures.que
3,
0,
10,
0,0,0,
**************** END OF FILE *********************************

DEVELOPMENT OF WINDOWS VERSION

For development of Windows version you should create graphical user interface which allow user to fill problem's description form and to view answers. You may look at Windows version including sources in the file 'gestures.zip'. It is 16-bit Delphi form.

For examle, code running GKM interface :
StrPCopy(PS,'gkmwin.exe ~eas-lg');
WinExec(PS,SW_MINIMIZE);

Set time delay :
{SET DELAY}
procedure TEASLG8.FormActivate(Sender: TObject);
begin
Timer1.Interval:=80;
end;

Verify if GKM interface processed input file :
{AFTER DELAY}
procedure TEASLG8.Timer1Timer(Sender: TObject);
var
ST: String;
I: Integer;
begin
if not FileExists('~eas-lg')=True then
begin
Timer1.Enabled:=False;
Timer1.Interval:=0;
ComboBox1.Enabled:=False;
AssignFile(FT,'~eas-lg.gkm');
Reset(FT);

This development of form is routine task, but takes significant amount of time. Form writes answers of user into a file with any name without extension (for example '~1') in a following format:

(First 8 lines from file 'gkm.ini' with a possible modification by user
(Maximum grades deviation etc.)):

Title of e-knowledge system
db.gkm
db2.gkm
Questionnaire file name (.que)
3,
Maximum grades deviation (%),
Minimum value index (%),
0,0,0,
(Problems description):
Problem's title
Sign1, Grade, Sign2, Grade,
0,

For this example:

**************** START OF FILE *********************************
Electronic Knowledge System: Language of gestures
db.gkm
db2.gkm
gestures.que
3,
0,
10,
0,0,0,
Problem description
1,,2,,
3,,5,,
0
**************** END OF FILE *********************************

After that form run 'gkm.exe' with file name as parameter, for example 'gkm.exe ~1'. Form must check for completion verifying presence of input file. Disappearing of input file from directory means finishing. After that form should read result file with same name as input file, but with extension '.gkm', for example '~1.gkm'.

**************** START OF FILE *********************************
General Knowledge Machine v4 e-knowledge base interface
Copyright (C) 1987-2002 Konstantin M Golubev
Web site http:/fast.to/gkm, Email gkmgeosite@yahoo.com
Distributed under Lesser GNU General Public License
Snb:2000 Enb:8000 Mex: 100 Min: 100 Mmm:1000 Snm: 64 Enm: 64
Language of gestures

Problem description
1, 1," 1.Nodding. "
2, 1," 2.Shaking one's head from side to side. "
3, 1," 3.Smile. "
5, 1," 5.Raised shoulders. "
0,0," "

======= HIGHLY VALUABLE PROPOSITIONS ==================================
PARAMETERS: Grades deviation = 0. %, Min. Index = 10. %
Index -- Number -------- Proposition --- (Total 6 of 99 ( 6. %))
63., 1,"1.Consent. "
63., 2,"2.Disagreement. "
63., 3,"3.Benevolence. "
63., 4,"4.Threat. "
33., 5,"5.Ignorance, incomprehension. "
22., 68,"68.Interest. "

============================================================= 1 ==
Signs
1, 1," 1.Nodding. "
Proposition 1
1.Consent.
-------------
Complete description
1.Nodding

============================================================= 2 ==
Signs
2, 1," 2.Shaking one's head from side to side. "
Proposition 2
2.Disagreement.
------------------
Complete description
2.Shaking one's head from side to side.

============================================================= 3 ==
Signs
3, 1," 3.Smile. "
Proposition 3
3.Benevolence.
-----------------
Complete description
3.Smile.

============================================================= 4 ==
Signs
3, 1," 3.Smile. "
Proposition 4
4.Threat.
------------
Complete description
3.Smile.

============================================================= 5 ==
Signs
5, 1," 5.Raised shoulders. "
Proposition 5
5.Ignorance, incomprehension.
--------------------------------
Complete description
4.Opened palms.
5.Raised shoulders.
6.Raised eyebrows.

============================================================= 68 ==
Signs
3, 1," 3.Smile. "
Proposition 6
68.Interest.
--------------
Complete description
3.Smile.
84.Look askance at partner.
85.Slightly raised eyebrows.

============================================================== ****** ==
**************** END OF FILE *********************************

Form must format result for good viewing by a user.

Please use e-mail address for consulting : gkm-ekp@users.sf.net

LICENSING

E-knowledge base tools are under Lesser GNU License and are available for download from gkm-ekp.sourceforge.net.

Best regards
Konstantin M Golubev
GKM Research Group coordinator