Paper 4

Paper Title:Toward a New Generation of Semantic Web Applications

d'Aquin, M.; Motta, E.; Sabou, M.; Angeletou, S.; Gridinoc, L.; Lopez, V.; Guidi, D.;
Intelligent Systems, IEEE Volume 23, Issue 3, May-June 2008 Page(s):20 – 28

Three Critical Questions

Monday

Group 1:

Member Name:Chiranjeevi Ashok Puvvula

 As we know that problems are not always simple to solve, there are more complex problems which can be treated as “high-order problems”. How can “Watson system” or the other search engines that are specified in the paper solve the complex logic associated problems?
 When we consider the “Limitation Problem”, even the results produced by the Watson system are not acceptable. So how far the information presented in figure-4 is acceptable?
 How can web crawlers overcome the problem of “privacy polices”? these privacy policy comes into picture because most of the web crawlers do web archiving in order the get the semantic data.
 How the problems of robustness, interactiveness combined with “KA” problem will be solved?

Group 2:

Member Name: Srikanth Kodali

1. The author says that, the problem of knowledge acquisition bottle neck can be eliminated by resource and knowledge sharing. But, the Domain knowledge is developed according to that domain’s requirement. Is it possible to integrate different knowledge sources developed for a particular domain requirement?

2. The article proposed that Watson semantic web is best solution for current semantic web requirements. Watson semantic web will take different levels of crawlers and different knowledge sources to filter the semantic web content. The time complexity to implement in real world scenarios is very high. Does this application complete this process in minimum amount of time which is feasible to real world scenarios?

3. Page ranking algorithm sorts out most visited or most popular web links. Watson method ranks the page according to the ontological relationship. The basic limitation of the Watson method is though it is ontologically accurate; it fails to satisfy the user’s requirement there by making whole system useless. How can this application give the accurate results to the end user?

4. Watson model deals with the ontological relationship between the ontologies. But how two contradicting ontologies are handled is not mentioned. Does wantson model able to deal with the heterogeneous ontologies?

Group 3:

Member Name: Sunil Kumar Garrepally

1. All the present semantic search engines sorts the results in accordance with the ranking, or popularity. But, the reliability, size, complexity and security aspects of these are not considered. Is it possible to make the semantic search engine to search considering these issues?
2. The triple data model makes it easy to accept requests and automate the responses. Even if we are saying that ontologies automate the data, we are following structured data models. Will they work even with the unstructured data as queries?
3. The results of the semantic web will be same for all the users. For example, the patient needs only what is the purpose of the drug, but for a doctor, it should show the chemical compositions even. Can the semantic web serves this purpose?
4. The existing web searching simply searches with keywords. But, semantic searching includes not only the keywords searching, searching synonyms and ontologies that may slow down the system. Is any research going on in improving the speed of semantic web?
5. As mentioned in the paper, it is possible to apply one domain knowledge base to the other domains. But, some of the key principles and laws in both the domains may vary and small variations may lead to larger problems. Is it a good idea to re-use other domain’s knowledge base in this case?

Group 4:

Member Name:Ramya Devabhakthuni , Nikhilesh Katakam

1)The paper stated that the WATSON gateway provides a single point of access and may provide inconsistencies and duplicate data. But I question how the consistency is achieved using this gateway when it involves large amounts of heterogeneous data.
2)The paper stated about many semantic web technologies and specified the limitations of each of these technologies thus proposing the other. But the solution for these limitations has not been discussed in the paper.
3)The semantic technologies may retrieve metadata and complex data which may contain sensitive information. How the factors like trust, quality and scalability is achieved in this case.

Group 5:

Member Name: Lokesh Reddy Gokul

• The author doesn’t mention the architecture of Watson API which may not be using new techniques as the author doesn’t mention any new techniques of data mining/semantic integration in the paper. So what special techniques/manipulations might have been used in it’s implementation?
• The author mentions the power-x versions of Magpie and Aqua where user can browse through the semantic information. This again brings is to the point that even in swoogle or google the user has to perform the search and the criterion in next generation semantic web application should be a minimized user interaction. How is this issue addressed?
• The author says that watson’s API ranks the ontologies in terms of quality rather than on popularity. what is the difference between the semantic quality and popularity of data according to the author?

Group 6:

Member Name: Palla Sangram

1)The author discussed about various aspects but he dint discussed about the security. How the security has been achieved in the new generation applications?
2) If the numbers of queries are not limited for user then several unnecessary results occur. How to avoid this problem without limiting the number of queries?
3) Developers should give more importance in defining the ontology which would finally give us a good ontology structure. How this is done?

Group 7:

Member Name: Priyanka Koneru & Mohana Siri Kamineni

1. “ Why is it not possible to transfer directly a domain’s expert knowledge to a machine? AI focuses on providing a relationship between system’s knowledge representation and it behavior but this suffers a unresolved problem called ‘ KA bottleneck’. How does ontologies and information sharing play a major role in the development of knowledge based agents and overcome the KA bottleneck problem? “
2. “The author didn’t mention the reasons why intelligence becomes a side effect of scale rather than of sophisticated logical reasoning when we move from classical ‘KBS’ to the ‘Semantic web applications’? “
3. “ A single knowledge source cannot provide sufficient information for the requires application. So one need to gather information for several sources. So when we collect data from several sources, data inconsistency and redundancy occurs. On what factors is the integration done ? what domains are to be considered while integrating arenot mentioned in the paper? “

Group 8:

Member Name:Durga Maheswari Muppalla

1.Semantic web as large scale knowledge opens up large amount of complex data. How the issues like trust and the quality are achieved when lot of complex content is shared.
2.How the problems like “duplicate ontologies” and the problems with the interconnection between the ontologies which results in the repeated results in the search can be solved.
3.It is stated that the ontologies “reflect different views” and the “derived structures” can suffer problems when they are used in “combination”. How the “inconsistencies” can be solved when deriving the knowledge from multiple ontologies.
4.How are the issues related to the “online ontologies” like the “online ontologies” having the poor coverage for the “tag types”, “multilingual terms” etc can be dealt?

Wednesday

Group 1:

Member Name:lattupalli,voruganti,pelluri

How can “TRUST” be the key factor in supporting applications to select resources and ontologies? Do you really think it would be reliable enough to depend upon trust?

Does the Knowledge acquisition bottleneck still persist now? If at all it is being handled now how reliable is it? the reason I ask this question is, the requirements of organizations from the semantic web increases rapidly which implies that there needs to be a lot of intelligence in the application, in such a case if at all there is any contradicting data it will eventually lead us back to the bottleneck.

The Watson still doesn’t have any advanced mechanism to efficiently discover fine-grained relations, as we go ahead into the next generation where a lot more knowledge is to be maintained and efficiently evaluated.So do you think the Watson will be reliable enough to detect the semantic relations between ontologies to aid the combinations?The only available mechanism is the simple duplicate detection problem which we know doesn’t use any logical functionality. Hence there will always be a question against this issue over the Watson

Group 2:

Member Name: Addagalla Satyanaga Manoj Kumar; Bobbili, Shrinath; Gopinath, Sreejith

1) The author mentions that providing smart applications that make use of some kind of data may be an incentive for people to share their own similar data. The question is whether without this data in the first place, would the applications be as effective as to act as an incentive?
2) In explaining that the KA bottleneck problem dictates that if the economics of acquiring and storing related data is costlier than the benefits derived out of the system, it probably is not worth the effort of building the system. The question is whether this is an important issue unique to the semantic web technology area; is it not true of any other field, databases for example?
3) The author discusses a few applications and concludes that they cannot exploit the large scale online knowledge. Is it positive to conclude that the semantic web applications cannot exploit large scale online knowledge by just considering a relatively small set of results?

Group 3:

Member Name: Swathi M Shastry

1) How can the size of the semantic web measured. There must be a common
measuring factor, a universally agreed upon notion of what constitutes
a semantic document.

2) How do the authors propose to design techniques for discovering
extension and compatibility of ontologies?

Group 4:

Member Name:Shaiv, karuna priya rameshwaram, anusha vunnam

1) As it is said that reducing barriers among formal definition is a real bonus, in what way they are going to reduce the barriers is not obvious in the paper.
2) More over we are going to develop ontologies and what we are gaining if no one use those ontologies , we don’t know exactly what will happen at this step.
3) Even though it was fine to make a system open resource in order to achieve one of the most features of semantic web , how could it would be possible? Whether all will agree to make their systems open and to be shared with others.

Group 5:

Member Name:

Group 6:

Member Name:

Group 7:

Member Name: Boda,Vamshidhar Reddy, Goyal Saurabh and De Morais Andrade,Pablo

• Only simple analysis tools have been used till now to study the large scale information. As size of data available keeps increasing, so Data Mining can be good option to analyze the data and get meaningful result?
• The analyzed aspects used did not able to capture the full quality of the knowledge. Author used Watson Semantic Search API in different web application to improve the trust and quality, but didn’t able to do so properly. How can author say that Watson semantic search provides better API?
• Author says providing single access point, so that all application can refer to this gateway. How Author is sure that this will analyze and provide the best result? It can also cause the bottleneck.

Group 8:

Member Name:Brugu Kumar Bhargava Mudumba

1) The Author discussed the differences between the semantic web applications and the traditional KBS and discussed several points but what in case of security?
Which is more effective?

2) The author mentions that the Watson method unlike sindice and swoogle doesnot limits the number of queries per user which suggests that there is definite possibility of finding the solution for the query but it also means that there are many outputs which are unnecessary. What are the steps to be taken to avoid this in future?

3) The target applications that exploit the heterogonous semantic web resources have to be addressed in 1st generation or KBS. Can this be done using the semantic web?

Group 9:

Member Name: Satish Bhat, Holly Vo

1. What is the human role, and how can human interactions be inserted to the infrastructure to improve data quality and trust?
2. Should relation discovery be distributed in groups of domains and then in layers of groups to improve performance?
3. Is performance of a real-world large-scale knowledge base acceptable? (now and in future) If it’s not, should we consider partially query result? And how?

Group 10:

Member Name: Sunae Shin, Hyungbae Park

1) They enumerate current problems and limitations of Semantic Web applications and showed how to overcome these problems and limitations. It is easy to understand why we need new mechanism for Semantic Web applications.
2) They don’t describe any disadvantages or limitations of the Watson Semantic Web gateway.
3) Watson processes and indexes online semantic information so that applications use these data in a lightweight fashion. I think this behavior of Watson needs enormous system resources. However, they don’t mention anything related to possibility and capacity of Watson’s system.
4) One of advantages of Watson’s API is no restriction of the amount of data (a user’s number of queries per day or the number of query results). I think they should mention about time complexity related to this procedure.
5) They describe what PowerAqua and Flor can do but don’t explain how they can do. It would be helpful if they put some explainations or details of mechanisms or algorithms for both of applications.
6) They assess the online ontologies’ information quality by evaluating 1,000 randomly selected mappings. This is a kind of sample inspection. Thus, they should explain the eason why their evaluation is correct.

Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License