Paper 5

Paper Title: Building a Pragmatic Semantic Web

Alani, H.; Hall, W.; O'Hara, K.; Shadbolt, N.; Szomszor, M.; Chandler, P.;
Intelligent Systems, IEEE, Volume 23, Issue 3, May-June 2008 Page(s):61 - 68

Three Critical Questions

Monday

Group 1:

Member Name:Abilash Amarthaluri, Bharadwaz Somavarapu

1. Author of this paper says that they have designed an approach to be practical, fast and cost effective which could be more pleasing for the organizations in a long term process. But what about the organizations which are working only on a short term basis and require the semantic web apllications to the maximum extent. How could the organizations build the required ontologies in a short span of time and also the budget? Is this not skeptical?

2. The author suggests building ontologies over the time, by scaling them to fit the individual information assets rather than building the entire domain.But how does the ontologies satisy all the specific needs? Any ontology is meant to satisfy all the requirements.Is it suggestable to build small ontologies which satisfy only individual assets?

3. The problem of misuse of data or information by other organisations once the Sematic Web technology is enabled access and reuse is adressed in the paper saying that We need to maintain a separate Knowledge Base with its own ontology by tranferring each resource into that KB. But would that not be a wastage maintaining a separate knowledge base inorder to provide security and privacy of data? Rather I suggest finding some better tools to resolve that problem.

Group 2:

Member Name: Sairam Kota

• First and foremost the author has to be appreciated for hitting the nail on the head, when he identified the private gain issue for organization and then touched it up using the pragmatic angle to technology adoption. This is crucial because adoption of any technology strongly rests on its ultimate return on investment value.
• Added to this factor are the 6 common misconceptions which the author lists in table. By using the two project examples the author clearly makes the case of semantic web strong.
• However the place where he misses the point is when he considers things only at an intra organization level, added to it when he stresses about financial sense of this whole idea, nowhere in this paper have they come up with the numbers saying how much the transition caused, how much money did it generate or in short ROI.
• While authors do talk about four main concepts which constitutes their central idea of pragmatism, they do fail to illustrate the importance of the provenance and privacy issue in any of the two projects. A few lines about the relevance of these issues would have greatly benefitted the paper.

Group 3:

Member Name: Sunil Kumar Garrepally, Yaswanth Kantamaneni

1. Semantic web technologies uses data sharing and integration concept that makes it widely used across the globe. But, it may causes security issues to the secure data of the organization and their policies. Are there any methods that could separate the secure data from the general data in the ontologies?

2. It is necessary to access data from multiple sources over the web. It requires mapping data that supports our domain. Will this conversion of the data into RDF format and mapping it increase the load on end system and makes the system slower?

3. The wide availability of data and its reuse may result in problems such as data misuse that makes the system nonviable. So, it is very important to maintain the signature / authorization in the data sets that are to be shared. Which digital signature algorithm would be suitable without affecting the ontology? Are there any laws defined for semantic data misuse as in the case of traditional web criminal laws?

4. As we are accessing the data from remote data bases which are of different quality, the quality of the integrated application can not be assured. Are there any methods which will determine the entire quality and reliability of the application?

Group 4:

Member Name:Ramya Devabhakthuni

• The paper stated that there exists inconsistency by retrieving data which is integrated from multiple data sets. But the elucidation of these duplicate sets is not being discussed.
• The projects were developed by restricting the size of the domain to small size. What happens when the ontology is to be extended or what if new data sets are to be integrated?
• Integration deals with combining large sets of data enabling reusability. How the problems like privacy and access to public data be achieved in this case?

Group 5:

Member Name: Lokesh Reddy Gokul, Ritesh Mengji

• The author says that in development of AKTivePSI, they have used Jena API while protégé is still the most used ontology API with a richer set available as perceived generally. A clear explanation of why such a scheme is used is not provided. Such a explanation is needed because the future world when trying to attempt to build their own applications would definitely look into this and question why protégé is not used. Is it because of application specific implementation or any other reason that the author forgot to mention or is it a possible intentional remiss?
• Also, author does mention that manual mapping has been more useful and easier in this particular scope and hence is preferred to the automatic mapping and after that, later in the paper he mentions about the inconsistencies and doesn’t give a clear explanation about the inconsistencies and the other problems which are left unresolved. Are such an independent decision made by the author’s team responsible for that or if not what is the reason for unclear presentation of inconsistencies and the reason for the use of other methodology than that is commonly perceived?
• The author mentions about the avoidance of disruption to the existing knowledge bases but that is an potential disaster for information access and updation as the main purpose of semantic web is common information sharing and updation and re-use. What could be a possible concrete solution model for such a problem as even avoiding disruption to the data the author says is a crucial step towards achieving a practical Semantic web?

Group 6:

1) For each and every application developed using semantic web technology needs two formats of data i.e., one for the human beings to understand and the other for the machine. This process is more time consuming in both private and public sectors. How this problem is going to be solved?
2) For holding the Knowledge base together RDF triples are generated. But the Syntax that the RDF frame work defines is hard to understand by the Humans since it is very much verbose. How to deal with this?
3) Using the semantic web technology from transforming from a medium scale to the large scale has some disadvantages like using a third party to Database etc. How this problem can be solved?

Group 7:

Member Name: Kishore Kumar Mannava & Mohana Siri Kamineni

1.“When each “organization” has large number of small and simple ontology’s. How can we achieve “mapping” of these large numbers of ontology’s present across various “organizations” across the world in a simple way?
2.What are the major steps that can be taken when construction of ontology’s using large “databases” with complex relations between the data without affecting the “database” performance?
3.When integrating the multiple sources to reduce the “knowledge gaps” across several organizations, how can ambiguous relations across various “KBs” be resolved? And how can the “organizations” overcome redundancies?”

Group 8:

Member Name:Durga Maheswari Muppalla

1.How the issues like the privacy can be achieved when two organizations are dealing with the highly confidential data as most of the “industrial and government circles” are new to the concept of presenting or “publishing” the data without controlling the presentation context.
2.In the aviation industry example much of the data is duplicated in different sources and stored in the terms of “instances and the concepts”. So this would seriously affect the granularity. In case of small organizations the data is mapped based on the granularity, how the mapping could be done as there are many “replications” done.
3.In order to reduce the complexity of the ontologies the scope and the domain are restricted. What if the particular domain has to add more information in it otherwise which leads to the loss of some useful information. How these restrictions for the scope and domain can be done without losing the useful information.

Wednesday

Group 1:

Member Name:voruganti,lattupalli,pelluri

As author implemented SW in two organizations, which have less domain information, so for him developing ontology may look easier. But what about the organizations which have larger domain? Developing Ontologies and mapping for them is no longer an easy work. The cost and work involved will be increased. So, companies may not intend to implement SW in their organizations.

Everywhere in his projects author implemented simple scripts to develop RDF triples. But can these simple scripts are capable of developing semantically correct triples?

Author has shown all the positive sides of the project. What measures are taken when erroneous results are achieved due to wrong mappings and weak scripts (which are used to produce triples).

Group 2:

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

1) The authors have mentioned some points that could encourage organizations to use semantic web technologies. While these are realistic, the one-size-fits-all approach does not work in most cases, since most organizations have a specific business model and processes; adopting a new technology could incur deep business knowledge and superior consulting skills.

2)The authors broadly mention some principles that they feel could reduce the pain for organizations that are adopting semantic web technologies. We think that these points are valid; but nevertheless, seems to be a case of “easier-said-than-done”. Each business is unique and how actually disruption to existing infrastructure could be minimized or well-focused technologies could be used, is still an open question. Again, each point could be relevant or irrelevant, painful or easy, depending on the organization.

3) The authors back their arguments by citing examples of a public domain application and another in the private domain, that they have built using the semantic web technology. The critical question that should be asked here is whether these applications are a true and sufficient representation of the real world, whether they can really justify for the wide variety of applications.

Group 3:

Member Name: Chi Zhang

1. In the paper the authors mentioned the by-product they discovered when integrating the data sets. But they didn’t state clear how they deal with it. If I were the designer, I would communicate with officials and other personnel who is in charge of the documents to make clear the details to the greatest extent in order to develop such a KB. If the problem can’t be solve in the development phase, how can they deal with it in the maintenance phase, since then it will be more costly to do so?

2. How did they obtain the database of these organizations? the author should add more details to that about how they contact with individual agencies other than stating the added-value, which I personally think will greatly benefit future research.

3. In the MROExpressway, why didn’t the author state clear how they connect different sources of data together? The author just mentioned they stored mass data together and customer could query the database. And they only provided a small example on coating the flight which in my point of view is not sufficient to support their words.

Group 4:

Member Name:Karunapiya Rameshwaram, Shaiv, Anusha Vunnam

Critical Questions:
1)The paper says that there are many benefits on using semantic web and that these benefits are different from that of known cost benefit. What categories do these benefits fall into?
2)One of the solutions the paper provides for not disturbing the structure of an already available database is to layer ontology on top of this database. What does this mean? Will this not lead to a complex structure?
3)If some ontology mappings lead to data inconsistencies then what measures are to be taken to avoid such a disaster? Should the ontology be entirely rebuilt?

Group 5:

Member Name: Rahul Mootha, Rahul Reddy

• In designing a semantic web application the no. of ontologies is an important issue. Then how far building a separate ontology for a resource is feasible though it benefits with a feature of security?
• Simple scripts across databases and ontologies of same kind can be reused to generate RDF s. Is it secure in reusing the same ontology?
• The author says that sw adoption costs are low and can be done by using simple scripts to convert data into rdf ‘s .He provides an alternative to achieve the low semantic cost and that he proposes by saying that an ontology can be built on top of a non rdf database . Will such an ontology be a true semantic web application?
• Integrating data sets may lead to inconsistencies , so how is the problem resolved ?

Group 6:

Member Name:

Group 7:

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

According to the paper, when questioned about the security and privacy of information. The response is that “… privacy is a complex issue … “ and “Many of us are not prepared to surrender our privacy for gains in efficiency or monetary benefit .. ”, others not. Many company definitely are not prepared to share their information on-line and might never be. So the Semantic Web wouldn't work for them? Only the sub semantic web, the “intra” semantic web?
• Most of the integration shown by the paper (integration of data sets) can be done, and have been successfully built today in many companied without the use of Semantic Web. Database integration is available for any business based Database System and corporative data information can be shared by the use of ERPs, so why to use web semantic?
- According to the paper the introduction of RDF format would bring advantages that actually were promised before when XML was introduction and any other data pattern format. So how many times will the companies accept and invest money on the introduction of new patterns? Couldn't ontology and semantic web re-utilize the current format, in this case XML?

Group 8:

Member Name: BHARGAV SANDEEP RAMAYANAM

1) The author suggests to minimize the disruption to the existing infrastructure but when a new knowledge is added it definitely leads to reengineering. How is the balance achieved when both are inversely proportional ?
2) Machine does less mistakes than humans. In practical can we build a automatic ontology mapper ?
3) The paper says that the technology is still in the developing stage then we need to make the human understand and also should make the system understand which is time consuming. Can this be avoided?

Group 9:

Member Name: Satish Bhat, Holly Vo

1. Knowledge based collaboration between companies is tough to achieve unless both of them gain something substantial from the purpose. Will public service serve as a motivation companies to go for shared profits by sharing KBs? In this competitive market are companies willing to compete together on the same page?
2. Sharing of KBs accompanies with it the problem of privacy. Privacy laws are strict. Who will monitor privacy laws for companies who have agreed to share KBS? In case of mal practice what will be the repercussions?
3. Who will do the mapping between knowledge based? Is it a automated process or manual process where domain experts are required?

Group 10:

Member Name: Sunae Shin, Hyungbae Park

1) They visited several organizations to investigate their opinions and thoughts regarding tradeoffs between costs and benefits of Semantic Web. They used these data to correct some common misconception about the Semantic Web. I’m not sure that they really did good researches for this. This information can be too subjective. For more accuracy, they need to explain at least how they chose those organizations and whom they met with.
2) They didn’t mention the way how to process a large number of the concepts and relationships for constructing ontologies. They just deal with affordable databases when they developed ontologies.
3) They found some inconsistencies when integrating data sets occurs. They said this is useful for the data provider since inconsistencies can represent insights into a data set’s quality.
4) The paper mentioned that undersized and plain ontology is better for reducing cost and time consumption. Connecting ontologies for information sharing might makes ontology larger and complex. They should explain the connecting method and how the linked ontologies can be stay in small size.

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