jump to navigation

Expert Systems in Conservation

Greek Version

Author email:

Introduction

This project was the MSc Thesis for a Computing Science cource at Birkbeck College, London and was presented at the 1999 AI*IA «Intelligenza Artificiale per i Beni Culturali» Workshop that was held at the University of Bologna. The domain is the conservation of small wrought iron objects found in land excavations. Since this happens to be one of the first expert systems for archaeological conservation I though someone might benefit from my experience if I publish the whole thing on the web. Most conservators are not familiar with the basic principals of expert systems and their possible application in conservation so I have also added this small introductory page to provide some basic help on these subjects. So if you are still interested keep reading, any comments are also appreciated. Feel free download it, but if you want to use it as part of your work I do require you make referance to the project, to me and my site.

Downloading and running the program

You can download the expert system by visiting this page and clicking the «Download Zip» button. To run the program, simply follow the User guide in page no.17 of the report (MScProject.doc). Although you might not be thrilled by the idea of a computing science project report I think is very important that you read the entire thing before you try to run the program.

What is an Expert System?

Expert systems appeared in the mid-seventies and so far have been the most successful branch of Artificial intelligence (AI). They are programs, which use non-numerical domain-specific knowledge to solve problems with the competence comparable to that of a human expert. The knowledge they contain can handle problems of a very specific nature and because it is non-numeric it is often not exact, in the same way that a human’s knowledge is imperfect. Expert systems are used when obtaining precise knowledge to determine a situation is very difficult or time consuming and reliable results can be obtained by following ‘rules of thumb’. Example of such a rule would be:

IF

the animal is warm blooded and purrs

THEN

the animal is a cat

It is not necessary to do a complete physiological analysis of the animal to determine if it’s a cat, a small number of carefully selected characteristics can give a relatively accurate result. The program simply requires the user to input the relevant information that will enable it to give an answer. In this case the program would reach its conclusion by asking the user a series of questions e.g. Is the animal warm blooded? Does it purr?

Expert systems can also contain the element of uncertainty, this means that they can arrive to a conclusion even when all the evidence to decisively prove the conclusion is not known. They can also explain there way they came to conclusion so that their reasoning can be checked and in some cases allow the user to explore multiple lines of ‘what if…’ type questions. Most of them also provide efficient mechanisms for adding, changing and deleting knowledge whenever necessary.

Advantages of expert systems

1) An expert system can be available 24 hours a day 365 days a year.

2) Well-designed expert systems are able to explicitly explain in detail the reasoning that led to a conclusion. Human expert may be too tired, unwilling or unable to do this all the time.

3) The expertise is permanent. Unlike human experts that may retire, quit or die the expert system’s knowledge will last indefinitely.

4) Depending on the software and hardware expert systems may respond faster than a human expert.

5) Expert systems contain the knowledge of many experts.

Disadvantages of expert systems

1) Expert systems can’t draw analogies from other sources to solve a newly encountered problem like a human would, in other words they can’t be creative.

2) Human experts automatically adapt to changing environments; expert systems must be explicitly updated.

3) Human experts have available to them a wide range of sensory experience; expert systems are currently dependent on symbolic input.

4) Although inexpensive to operate expert systems are expensive to develop and maintain.

Expert systems have been used successfully in interpreting mass spectrograms to identify chemical constituents (DENDRAL), analyze geologic data for minerals (PROSPECTOR) and for the optimum configuration of computer systems (XCON/R1).

Conservation expert systems

In resent years there has been some work on the possible applications of computers in conservation. These attempts tend to concentrate mainly in the fields of digital imaging (e.g. VASARI Project) collection management databases and programs for processing data from instrument measurements. I believe that expert systems combined with data provided by the applications described above can offer considerable assistance to the conservator.

There are at least two possible ways of using expert systems in conservation, 1) for determining the condition of the object 2) for choosing suitable conservation treatments. My research was concentrated mainly on the possible use of expert systems in determining the optimum conservation treatment for specific types of artifacts.

When having to choose between conservation treatments for an object, the conservator has very little time and resources to acquire details and must decide the course of action based on partial data. A treatment expert system would aim to aid the conservator’s decision by suggesting the best possible treatment based on the condition of the object and other types of partial information. It could also suggest to the user some alternative-but perhaps less suitable- treatments in case restrictions in time and resources prohibit the application of the best treatment, and give the possible advantages and disadvantages between them. It should also explain the reasoning behind every answer. These expert systems will have to be developed initially for very restricted domains (e.g. small wrought iron objects found in land excavations) and could become later integrated in larger more general-purpose systems.

Such an expert system could also be useful as an education tool, something that can be used to test a student’s knowledge on a specific subject, comparing his decisions with that of the system in a hypothetical treatment case. Also the process of developing an expert system itself has an indirect benefit. It forces the knowledge of a specific domain (e.g. conservation of archaeological iron) to be put in an explicit form. In this way it can be examined for its correctness, consistency and completeness.

Before beginning research in the development of the actual system I had a few discussions with conservators about the possible attitudes towards such a program if it ever became available. On the whole most of them were positive but some of them stressed some objections about it.

The most frequently encountered objections were:

  1. Most conservation labs lack far more basic equipment and materials and must be able to satisfy these more urgent needs before acquiring a computer.
  2. The complexity of the treatment problems which are unique for each object and cannot be solved by something as restricted as a computer program. A conservator can easily deal with most problems without any advice computer generated or otherwise.
  3. In this way the conservator is being reduced to a technician that simply follows the instructions of a machine.

I feel that it is important that these objections should be answered.

  1. It is true that the budget of most conservation labs prevents them sometimes from obtaining even some of the basic materials and equipment and turns computers in an unaffordable luxury. But the price of computers is constantly falling and soon they will become an integral part of every day life. Already computers are beginning to appear in conservation labs for general administration purposes and record keeping. It is doubtful that in five years time you will be able to find a conservation lab without a computer.
  2. It is also true that some conservation problems are genuinely unique and require originality, very careful analysis and considerable skill to solve. But most of the problems a conservator might come upon have been encountered before and usually a solution or solutions exists for them. Expert systems are very good at suggesting solutions for such ‘trivial’ problems; this will hopefully allow time for the conservator to concentrate in dealing with the difficult problems an expert system can’t solve. As for the claim that every conservator should be able to deal with the ‘trivial’ problems without any help, I must say it is doubtful if even the specialists feel confident when dealing with ‘trivial’ problems. In any case a second opinion is always helpful and is usually very difficult to find.
  3. The job of expert systems is not to replace the human expert; it is to provide a second opinion. The ultimate decision belongs to the conservator, which is why it is necessary for the expert system to provide an explanation for its reasoning. In this way the conservator can judge for himself whether to follow the program’s advice or try to think of something else. On the other hand it is important the conservator has the necessary training/experience to be able to interpret the available information for himself.

I believe that a possible problem with expert systems in conservation might be the same with that of expert systems in archaeology. There is a considerable number of archaeological expert systems out there. Few people have heard of them and even fewer people have used them. This is not because they are useless but because no one seems to be willing to use them. Most archaeologists seem to feel uncomfortable with computers and try avoiding them as much as possible. As a result most archaeological expert systems, if not all, have remained in the experimental stage and have never been actually used in the field.

The same could happen to a conservation expert system even if it proves really useful.

A Note about this Project

As a computing Science project it concentrates mainly on the programming and design of the expert system. Anything that has to do with archaeological conservation is represented in a way that someone who is a complete novice on the subject would be able to understand it.

The entire project (research, programming and write-up) took about three months. The rules where extracted mainly from the relevant literature and a few interviews with the conservation departments of the Museum of London, the British Museum and English Heritage. Unfortunately it seems that it would be very difficult to develop a working expert systems for conservation treatment (at least of archaeological iron). Although expert systems are meant to work with relatively inaccurate information they require this information to be consistent and to have been obtained trough some sort of standard procedure. It was never the intention for this project to produce a working expert system but I hoped that it could demonstrate some of its capabilities. But instead of providing the user with the best treatment for each object, the system simply offers all acceptable treatments and excludes those that are completely unacceptable. The information acquired from relevant books and articles was too contradicting and inefficient to allow a real comparison between the treatments. Still, its results are more or less decent, they can be supported by the relevant literature and the system arrives to its-not very impressive-conclusion in the amount of time an expert would require for the same quality of answer. (Please read the Conclusions Chapter in the report).

The information from the relevant literature indicates that the long-term and short-term effects and even the comparative efficiency for most conservation treatments for land archaeological iron is actually unknown. With a few exceptions the research that exists on the subject is rather inconsistent and lacks any real evidence. (Please read the Conclusions Chapter in the report). Before any attempt is made on the creation of a working expert system it is very important that this type of conservation research is advanced so that the experimental data become reliable and meaningful. It has been suggested that statistical methods used in other fields like epidemiology might help in treatment efficiency research. The conservation department of the Museum of London has begun research on this field by trying to modify and apply some of these methods in conservation to determine the efficiency and short and long-term effects of conservation techniques on various groups of types of objects made of archaeological leather.

Even though the development of a conservation treatment expert system is still a long way, it is possible that expert systems will be much more useful in the field of conservation analysis where the information available is much more consistent and well researched.

Contact

If you want more information about this project you can contact me though email (lliberopoulouATSYMBOLgmail.com)

Bibliography

Expert systems

Introduction to Expert Systems

Giarratano, J. and Riley ,G. Expert Systems: Principles and Programming Second Edition, PWS Publishing Company, Boston 1994

CLIPS WWW Page

http://clipsrules.sourceforge.net/

Expert Systems in Archaeology

Coriosolite Expert System.

http://www.writer2001.com/exp0002.htm

Expert Systems for Lithic Analysis

http://www.hf.uio.no/iakk/roger/lithic/expsys.html

J. Huggett and K. Baker 1985 «The computerised archaeologist: the development of Expert Systems», Science and Archaeology 27, 3-5.

J. Huggett 1985 «Expert systems in archaeology», in J. Richards and M. Cooper (eds.) Current Issues in Archaeological Computing (British Archaeological Reports, Oxford), pp. 123-142.

J.-C. Gardin Artificial Intelligence and Expert Systems: Case Studies in the Knowledge Domain of Archaeology, Prentice Hall (Sd) 1989

Wilcock, J.D. 1985. «A review of expert systems, their shortcomings, and their possible applications in archaeology», Computer Applications in Archaeology 1985, University of London, 139-144

Wilcock, J.D. 1990. «A critique of expert systems, and their past and present use in archaeology». In Ennals, R. & J.-C. Gardin (eds), Interpretation in the humanities: perspectives from artificial intelligence, LIR Rep. 71, British Library, 130-142

In Computer Applications and Quantitative Methods in Archaeology annual conference proceedings

Conservation

For all things related to Archaeological/Art conservation (People, Organizations, Studies, Education and Training, Articles, Events, Mailing Lists etc) go to the Conservation On Line Site (CoOL)

http://palimpsest.stanford.edu

Σχόλια»

1. FarCry - Σεπτεμβρίου 25, 2006

Na po oti sta ellinika onomazontai eufii sistimata giati sti greek version sinexos os expert systems anaferontai😛

2. George - Αυγούστου 24, 2007

how can i download far cry

3. George - Αυγούστου 24, 2007

How can i download far-cry for x-box 360;

4. Lida - Αυγούστου 24, 2007

Get a bittorrent client and try to find it here by puting «far cry xbox» in the search field.


Σχολιάστε

Εισάγετε τα παρακάτω στοιχεία ή επιλέξτε ένα εικονίδιο για να συνδεθείτε:

Λογότυπο WordPress.com

Σχολιάζετε χρησιμοποιώντας τον λογαριασμό WordPress.com. Αποσύνδεση / Αλλαγή )

Φωτογραφία Twitter

Σχολιάζετε χρησιμοποιώντας τον λογαριασμό Twitter. Αποσύνδεση / Αλλαγή )

Φωτογραφία Facebook

Σχολιάζετε χρησιμοποιώντας τον λογαριασμό Facebook. Αποσύνδεση / Αλλαγή )

Φωτογραφία Google+

Σχολιάζετε χρησιμοποιώντας τον λογαριασμό Google+. Αποσύνδεση / Αλλαγή )

Σύνδεση με %s

Αρέσει σε %d bloggers: