September 27, 2008
Online Salon (Sep 16, 2008)
The 3rd Online Salon was held at 13:00 (GMT±0) on Sep 16 (Tue), 2008.
==================================================
Note for the salon
(Written by Ms. Hebin Lin, participant of the salon)
==================================================
---------------------------------------------------------------------------
General Information
---------------------------------------------------------------------------
Guest: Dr. Munehiko SASAJIMA
Assistant Professor, Dept. of Knowledge Systems, ISIR, Osaka Univ., Japan.
http://www.zukan.tv/2007/06/18/munehiko-sasajima/
http://www.ei.sanken.osaka-u.ac.jp/members/msasa/index-e.html
Facilitator: Dr. Kuniyoshi SHIMIZU
Assistant Professor, Dept. of Forest and Forest Products Sciences, Kyushu Univ., Japan
http://hyoka.ofc.kyushu-u.ac.jp/search/details/K001655/english.html
http://www.zukan.tv/2007/01/26/kuniyoshi-shimizu/ (Japanese)
Participants
6 people
---------------------------------------------------------------------------
Questions/comments from the participants to Sasajima (guest)
---------------------------------------------------------------------------
1)
Question:
Is "ontology" modeling based on empirical rules and probabilistic methods of human behavior?
Answer:
Yes and no. we can develop based on both empirical rules and probabilistic methods. In that way, it is yes. But it is not the main method. So, it is also no. We are looking at the actual users in the real world. We also borrow the idea from the field of philosophy. In the long term, what elements of this world were concerned is the concern. Currently, in ontology, there are guidelines, but still it is not sure.
2)
Question:
Is it allowed to introduce some irregular (or low probability) tasks intentionally/randomly in the modeling?
Answer:
Yes. As I mentioned before, currently there is only guideline. There is not strict rule. I would like to mention that it depends on what application and such introduction is good or not. Currently I apply ontology in analyzing user activity in mobile service. For example, mobile users involved in crimes, they may call for help in low probability. Or people suddenly become sick. In such cases, we should consider these users for low probability actions. To make the usability, we need to introduce such task, even though they are low in probability. Basically, ontology should focus on essential things. If they make the application better, I would suggest. In other cases, it depends on designers.
3)
Question:
Since it is supposed to be prioritized for high probability tasks, what should be considered in implementing the designs, such as costs and contributional effects?
Answer:
Interesting question. the first priority is ontology’s potential, and which question should be solved by ontology knowledge. But don’t pay attention to building costs. Many concepts should be modified according to the fields. In education, there are 2000-3000 concepts, in medicine, there are 100,000, in environment, there are about 2000-3000 concepts.
4)
Question:
How to evaluate the responsibility of the ontology made by people or designers?
Answer:
Frequently asked question. Many ways can be used to evaluate. By testing the usability in real practice, with suggestions we modify our model. In case of sharing ideas, we also consider reasonable comments. In medical field, there is common sense of terminologies, they already share the terms so that they could check by their views and definitions. We can also ask to experts and use their judgements.
5)
Question:
How about quantitative evaluation? Is it possible?
Answer:
In some fields, it is possible, for example, in medical field. The numbers of concepts are about 100,000. If to define only 7 terms, then it becomes a quantitative evaluation. Another way is using evaluation software. Such as speed to find the solution. We also evaluated the time to reach appropriate solutions.
6)
Question:
If so, there is possibility that this modeling will be used for advertising business. It may be abused, though.
Ansewer:
We also think so, that it can be used to advertising. If you have a good idea, we would like to listen to. We already model the activity and situations. For train timetable for example, we will tag the information then we could link the needs of the users to this information. Or in newspaper page, if people are looking into travel plans, then we can link such article by looking for the tags. We do have ideas to apply the models for advertisement. Currently, as you know, newspapers are providing services. Sometimes the linked advertisements are relevant but sometimes are not. They may be only for title but not for contents. I think that if we use in a way more semantic, we could connect them more precisely. In search engine optimization (SEO), for those which are not improving usability but only for business, the quality may not be good and it turns out to show the big companies. Our current research project is focusing on designing menu and analysis of user behaviour.
7)
Question:
Some attractive irregular tasks or actions bring people into a certain direction. That is called “misdirection”, which is a technique of magic. I just imagined that ontology could bring people into a certain direction like a “misdirection” magic. (e.g. advertisements, websites)
Answer:
It is our idea to direct to proper services. Since we do our research with docomo, we intend to improve the Icon service. If there is misdirection, docomo will lose reputation to the consumers. Services are not organized for the people. We analyze the user situation and user just select through the movement with their intentions (such as shopping).
8)
Question:
How do you test models you made? Do you provide some websites for users in order to get date from them? Or mainly collaborate with companies?
Answer:
We first test form the ontology view. For example, whether the definition is correctly defined. Then, we test from the expert view.
9)
Question:
In what way will ontology change the current situation of the Web and of the world? For example, to the direction to increase the diversity of the world or to make the world more “homogeneous”?
Answer:
Deep and interesting question. Some people consider their own world, diversity of providing information is not ensured. We are not aiming at homogenous. But we also use concepts and semantics of their world. Even in terms of a simple concept web, there are many understandings. With just text, some people can’t understand expert’s intention. But with contents, expert ideas could be used. For example, for environmental field, if ideas are informed through semantics, I could find out some ideas to contribute by using computers, because the expert ideas and knowledge are provided with contents and meanings and usages.
10)
Question:
When you list up the problems in your project, maybe you should be careful to list as much as possible. How do you manage to do it?
Answer:
We use information science to define user behaviour more details. In the beginning we start from abstract phrase. Then we decompose it into smaller size of situations. For these situations, we consider the problematic situations. For more and more problematic situations, we further decompose.
11)
Question:
Do you list them with a few people or do it alone?
Answer:
Firstly we list with few people then review it in front of many people.
12)
Question:
Ontology is based on domain knowledge, how to model the domain knowledge?
Answer:
There are two types of knowledge, domain knowledge and activity knowledge. First of all, we model activity as tasks but not consider domain contents. Then we model relevant things in the world. Secondly we describe domain things. So that we won’t get lost with domain knowledge since there are many ways to define domain knowledge.
13)
Question:
Can the ontology modelling add/remove/change the task lists and the task order correspondingly to the place where I am?
Answer:
Yes. For example, for mobile phone, we can arrange the priority of the task services by referring to the services.
14)
Question:
Do you have any advice on how to use ontology in our daily life?
Answer:
Usability is not so good right now. I want to help by improving the interface and mobile menu. To list up the problems in daily life, we give some support ideas for that situation. (Idea by Kachi: I think the way of thought is useful for problem finding and solving, something like MECE or some tools developed by consulting firms) For example, we apply to docomo, and people tend to see user activity very easy and don’t care so much definition but sometimes the sharing of ideas is critical.
---------------------------------------------------------------------------
Questions from the guest to the participants
---------------------------------------------------------------------------
1)
Question:
What do you expect from the technology of "Ontology?" What do you think would it be used in the future?
Answer:
I think Ontology is very powerful as a method in framing concepts and relations. Apart from the use based on engineering field, I wish it could be more integrated with social science in decision making, such as policy design. It seems to me that there will be potential for Ontology to be favoured by emerging new science or multi-disciplinary research, since it is useful in discovering possible scenarios and risks, and accordingly providing options and paths.
Re-Comment by Sasajima:
We have several projects and many of them are multidisciplinary research. Those projects are related to academic fields, such as policy making. One project is on sustainability science. People ask for ontology to define which field is related to other fields or other problems or situations. People are sharing their expertises. Visualizing relationships are developed. So people from different fields can discuss and share knowledge and problems. We aim at more people make outputs and share their ideas to make situation better and produce more knowledge. Currently, people are seeking for how to apply ontology to share knowledge. If you are interested in using ontology, please send me email and describe the domain. (Idea by Kachi: Ontology is a good tool to visualize problems in any field.)
