[ieee 2012 international conference on advances in social networks analysis and mining (asonam 2012)...

6
Leveraging Social Networks to Improve Service Selection in Workflow Composition John McDowall #1 , Larry Kerschberg #2 # Computer Science Department, George Mason University MS4A5, 4400 University Drive, Fairfax, VA, USA 1 [email protected] 3 [email protected] Abstract— In recent years, social media has expanded from a niche application with a student-focused user base to a mainstream tool used by individuals and business to maintain and expand their social and business networks. The result is a rich source of data that can help service consumers and service providers connect with each other in new and interesting ways. This paper describes research that leverages social media to help users optimize service selections when composing executable workflows from among both web services and physical services. I. INTRODUCTION Since the emergence of the Service Oriented Architecture (SOA) paradigm nearly a decade ago, a major attraction of the SOA design philosophy has been the possibility of dynamically composing available services into executable workflows. For a variety of reasons discussed in detail in [1– 4], this vision has not yet been realized. In the course of researching methods for enabling the dynamic composition of services, we came to the realization that the web service composition problem is a special case of a more general service composition problem [5]. Modern society is structured around a world of services, where specialists ranging from doctors and laboratory technicians to electricians and plumbers offer specialized services to any consumer who either cannot or does not care to perform those tasks themselves. In a globally connected world, service providers and consumers are no longer restricted to doing business within their immediate geographic area. By the same token, service consumers have an array of choices unparalleled in history. A homebuilder in Brazil can order custom ceramic tile from a small tile maker in Italy; an emergency room physician in a small town in Nebraska can collaborate with a radiologist in Canberra to interpret an X-ray. Improving the ability of service consumers to find providers whose offerings meet the consumer’s needs is as applicable in the physical world as it is in information systems. In this paper, we discuss our research regarding the applicability of social networks to the problem of matching consumers and providers in a service system. Building on the services science concepts described in [6], we propose leveraging existing social media and business registries to improve matchmaking among consumers and producers. The remainder of this paper is organized as follows: In Section II we illustrate the issue with a sample problem; in Section III we explain the relevance of social networks in service selection; in Section IV we explain our approach to using social media to developing service recommendations, In Section V we describe ways to enhance the utility of social media and business networks. We discuss our conclusions in Section VI and describe areas for further research in Section VII. II. SAMPLE PROBLEM Consider the simple example workflow shown in Fig. 1, depicting the steps that must be completed to purchase and install a light fixture. Fig. 1 Sample workflow for selecting and install a light To complete this workflow, the user must select a light fixture perhaps from a catalogue, in a showroom, or via a web-based shopping service. Once selected, the light must be 2012 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining 978-0-7695-4799-2/12 $26.00 © 2012 IEEE DOI 10.1109/ASONAM.2012.220 1310 2012 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining 978-0-7695-4799-2/12 $26.00 © 2012 IEEE DOI 10.1109/ASONAM.2012.220 1278

Upload: l

Post on 14-Apr-2017

215 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: [IEEE 2012 International Conference on Advances in Social Networks Analysis and Mining (ASONAM 2012) - Istanbul (2012.08.26-2012.08.29)] 2012 IEEE/ACM International Conference on Advances

Leveraging Social Networks to Improve Service Selection in Workflow Composition

John McDowall#1, Larry Kerschberg#2 #Computer Science Department, George Mason University

MS4A5, 4400 University Drive, Fairfax, VA, USA [email protected] [email protected]

Abstract— In recent years, social media has expanded from a niche application with a student-focused user base to a mainstream tool used by individuals and business to maintain and expand their social and business networks. The result is a rich source of data that can help service consumers and service providers connect with each other in new and interesting ways. This paper describes research that leverages social media to help users optimize service selections when composing executable workflows from among both web services and physical services.

I. INTRODUCTION Since the emergence of the Service Oriented Architecture

(SOA) paradigm nearly a decade ago, a major attraction of the SOA design philosophy has been the possibility of dynamically composing available services into executable workflows. For a variety of reasons discussed in detail in [1–4], this vision has not yet been realized.

In the course of researching methods for enabling the dynamic composition of services, we came to the realization that the web service composition problem is a special case of a more general service composition problem [5]. Modern society is structured around a world of services, where specialists ranging from doctors and laboratory technicians to electricians and plumbers offer specialized services to any consumer who either cannot or does not care to perform those tasks themselves.

In a globally connected world, service providers and consumers are no longer restricted to doing business within their immediate geographic area. By the same token, service consumers have an array of choices unparalleled in history. A homebuilder in Brazil can order custom ceramic tile from a small tile maker in Italy; an emergency room physician in a small town in Nebraska can collaborate with a radiologist in Canberra to interpret an X-ray. Improving the ability of service consumers to find providers whose offerings meet the consumer’s needs is as applicable in the physical world as it is in information systems.

In this paper, we discuss our research regarding the applicability of social networks to the problem of matching consumers and providers in a service system. Building on the services science concepts described in [6], we propose leveraging existing social media and business registries to improve matchmaking among consumers and producers. The remainder of this paper is organized as follows: In Section II we illustrate the issue with a sample problem; in Section III we explain the relevance of social networks in service

selection; in Section IV we explain our approach to using social media to developing service recommendations, In Section V we describe ways to enhance the utility of social media and business networks. We discuss our conclusions in Section VI and describe areas for further research in Section VII.

II. SAMPLE PROBLEM Consider the simple example workflow shown in Fig. 1,

depicting the steps that must be completed to purchase and install a light fixture.

Fig. 1 Sample workflow for selecting and install a light

To complete this workflow, the user must select a light

fixture perhaps from a catalogue, in a showroom, or via a web-based shopping service. Once selected, the light must be

2012 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining

978-0-7695-4799-2/12 $26.00 © 2012 IEEE

DOI 10.1109/ASONAM.2012.220

1310

2012 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining

978-0-7695-4799-2/12 $26.00 © 2012 IEEE

DOI 10.1109/ASONAM.2012.220

1278

Page 2: [IEEE 2012 International Conference on Advances in Social Networks Analysis and Mining (ASONAM 2012) - Istanbul (2012.08.26-2012.08.29)] 2012 IEEE/ACM International Conference on Advances

shipped and payed for, and then the user must find an electrician capable of installing the fixture, and must schedule the installation.

As one example of a service selection, the user may select an electrician based on any or all of several criteria: the recommendations of friends, reviews of different electricians’ work posted to web sites or by random selection from a business registry.

III. RELEVANCE OF SOCIAL NETWORKS Long before the era of MySpace and Facebook, social

scientists recognized the importance of social networks in economics in general [7] and as a factor in service use in particular [8]. These networks can be leveraged for finding services and for recommending the selection of a particular service provider from among those available.

A. Service Discovery The development of the Internet and its plethora of search

engines gives us access to more information than ever before in human history. With just a few minutes’ work, even a novice user can find hundreds of potential providers of almost any desired service. Many of the major search engines, as well as specialized business directories, offer location-based capabilities that let us search for service providers in a specific geographic area.

But having all this information at our fingertips doesn’t always help us find exactly what we’re looking for. Even the most sophisticated applications still have trouble deciphering the semantics of ambiguous terms [9]. If a consumer is looking for an electrician to install a light fixture, a search for “residential electrician” will return many potential service providers who specialize in wiring new houses but are not willing to take on small jobs like installing a single light fixture. Alternatively, a search for “handyman services” will return many potential providers who offer everything except electrical services.

The surest way to find a provider of the service we are actually looking for is often the easiest: just ask a friend or family member if they know a provider of that service. If they do not know, they will immediately return a negative answer; if they do know they will return an accurate answer. Whether we realize it or not, most of us already use our social networks for service discovery.

B. Service Recommendation Once a service consumer finds the potential providers who

offer the desired service, the consumer must select from the available providers to find one most likely to live up to the terms of the agreement and provide the agreed upon service at the agreed upon price. As in the case of service discovery, we often leverage our social networks for service recommendations.

Humans are naturally social creatures, and we place a great deal of value on the recommendations of friends and colleagues, especially when we are operating outside our own area of expertise. If we need an electrician, we will naturally ask a friend in the building trades whether a particular

electrician should be hired based on the belief that this friend will know from experience which electricians in the local area do quality work at an acceptable price. By the same token, if a businessman needs to hire a personal accountant, he may ask his business accountant for a recommendation on the assumption that the business accountant’s expertise in the field makes him more likely to know which of his colleagues in the personal accounting field will do the best job.

Based on our analysis of the available recommenders within social networks, we divide potential recommenders into five categories, listed beginning with the most trusted.

1) Friends: These are people we know and trust. Based on years of personal experience with them in a variety of situations, we develop an appreciation of their judgment that leads us to trust their opinion when selecting a service provider. It should be noted that not all friends are equally valid recommenders for all types of services; a friend who is a plumber may be ideal for recommending an electrician, but may not be as reliable as a friend who is a nurse when selecting a doctor.

2) Trusted Reviewers: These are people or organizations we don’t know personally, but whose opinions have proven accurate enough over time to accord them a level of trust with respect to their service recommendations within a particular field. Organizations such as J. D. Power and Associates have built businesses around the idea of being a trusted reviewer of services and service providers; their recommendation carries a lot of weight with many potential customers. Trusted reviewers can also be individuals whose opinions have proven useful over time. Web sites such as Yelp allow users to write reviews of businesses, and a given user’s reviews may prove accurate enough over time that a potential consumer may consider their opinion almost as trustworthy as a recommendation from a personal friend.

3) Trusted Social Media Pages: An example of this type of recommender is the Facebook page of a business where the potential consumer “Likes” the page (i.e., makes a positive assertion about the business, usually based on some personal knowledge of or interaction with that business), and therefore trusts recommendations made by that business. In this case, the recommender has proven trustworthy in some endeavour, and the provider’s recommendation of a fellow service provider carries weight based on the principle that the recommender wants to protect their own reputation and will not recommend service providers that it does not want to be associated with.

4) Trusted Sources: These are recommenders that potential consumers trust based on some criteria short of personal interaction with the recommender. This can range from businesses the consumer respects to celebrity endorsements. For example, if the local hospital uses a particular cleaning company, we might conclude that the cleaning company must be good. Alternatively, many people buy goods and services based on celebrity endorsements, on

13111279

Page 3: [IEEE 2012 International Conference on Advances in Social Networks Analysis and Mining (ASONAM 2012) - Istanbul (2012.08.26-2012.08.29)] 2012 IEEE/ACM International Conference on Advances

the assumption that a celebrity won’t jeopardize their own reputation by endorsing untrustworthy businesses.

5) Crowdsourcing: One of the advantages of modern communications technology is that it makes it possible to quickly aggregate the opinions of many people from many places. This can be done fairly easily, with the result that relatively small numbers of people (perhaps only a few hundred out of the millions in a given city), with no other connection to each other, can form the basis for recommending a service provider. If we know nothing about either of two potential service providers, knowing that 12 people recommended Provider A and 342 people recommended Provider B may be enough of a recommendation to support the consumer’s selection of Provider B.

IV. LEVERAGING SOCIAL MEDIA The emergence of social media such as Facebook, Google+,

and similar applications has broadened the scope of many peoples’ daily interactions beyond the immediate circle of family and close friends that live and work in the same geographic area. Social media makes it easy to stay abreast of the events in others’ lives and to communicate with distant friends and relatives easily and inexpensively.

Based on our research, we believe the online communities formed by social media and related applications offer an opportunity for both service providers and service consumers to develop an informal, decentralized online marketplace where consumers can find service providers and select the service providers that best meet their needs based on the recommendations of their social network.

A. Social Media Many businesses have begun embracing Facebook,

Google+, and other social media as an advertising medium and as a way to communicate with their customers. Businesses ranging from electrical wholesalers to restaurants caret Facebook pages where they can advertise on a global platform for little or no cost beyond the time spent periodically updating their page. For the sake of simplicity, and because of their immense share of the social media market, we our research in this area is concentrated on Facebook, although the principles described here are equally applicable to all social media.

Given that the barrier to entry is so low, the potential return on investment to any business building a Facebook page is high. Creating a page is free, and updates require only a few minutes, in contrast to the extensive planning and graphic design work that are part of a typical print advertising campaign. With the investment of a couple hours’ time for the initial setup, and a few minute’s time per week for updates, any business, an individual artisan can have an advertising reach that was previously only available to large multinational corporations. Almost any business could benefit by embracing social media and leveraging its power to build up their brand awareness.

A business page on Facebook can have benefits beyond just the page itself and the updates the business posts to the page. Many businesses find it useful to place “Like us on Facebook” widgets on their web sites, so users can click a single button and register a “Like” for that business’s Facebook page. That user’s friends will see an update about that activity, exposing the business to potential customers who may not have heard of the business. This simple person-to-person connection also serves as an implicit recommendation of a service provider by each user who “Likes” that provider.

B. Business Registries Business registries and web sites that specialize in business

reviews aren’t normally thought of as social media, but many of them have social aspects that parallel other social media applications and that may be of interest to service providers. There are several options available for users to search for and evaluate service providers, such as Craig’s List or Angie’s List. Many of these sites have no publicly available interface, making automated searching and evaluation difficult. Therefore, we concentrate our analysis on those sites that have public interfaces that can be incorporated into a service search and evaluation system. We discuss some representative examples below.

1) Yelp: Yelp is a web site specializing in business reviews. Their focus is on location-based reviews; that is, they assume the user is looking for reviews of local businesses or businesses within a particular geographic area. Yelp automatically populates basic information about businesses and allows those businesses to claim their entry and enhance it. However, the primary function of the entry for a business is to allow Yelp’s user base to rate the business on a scale of one to five stars (five being the highest rating); a sample is show in Fig. 2

Fig. 2 Sample Yelp rating

In addition to its business ratings, Yelp has social media

aspects that offer several possibilities for use as a source of service recommendations. Each registered review can create a profile that consumers can see, giving users some insight into the reviewers’ perspective. Additionally, the Yelp site and programming interface allow users to see any reviewer’s

13121280

Page 4: [IEEE 2012 International Conference on Advances in Social Networks Analysis and Mining (ASONAM 2012) - Istanbul (2012.08.26-2012.08.29)] 2012 IEEE/ACM International Conference on Advances

grading history so they can put each reviewer’s recommendation in context. For instance, one reviewer may give an average of two stars to businesses she reviews, so her five-star rating of a service provider may carry more weight than the same grade by a reviewer whose average grade is four stars.

Users and service recommender systems can take advantage of this insight into reviewers’ history to better refine recommendations by weighting different reviewers’ ratings based on their average rating and the value the user accords the reviewer’s opinion as well as their history of rating similar service providers. For instance, say a trusted reviewer whose previous reviewing history is focused on restaurants gives a low score to a plumber. This reviewer’s recommendation may be accorded a lower value than a mediocre score from a unknown reviewer whose reviews are focused on home improvements.

2) eBay: eBay is widely known as an auction site, where users offer goods for sale to the highest bidder or to anyone willing to pay the “buy now” price. Less well known is eBay’s “Specialty Services” section, where service providers can offer their services to eBay users. The Services area on eBay is thinly populated at the present time; our review of the available services revealed only a few service providers, largely focused on the area of graphic design.

One potentially useful aspect of eBay as a platform for service discovery is that, unlike most sites that offer access to businesses, eBay includes the capability to rate buyers as well as sellers. In any business transaction, both the buyer and seller are assuming some level of risk: The consumer runs the risk that the service provider will not provide an acceptable service and the service provider runs the risk that the consumer will not pay for the services consumed. Leveraging eBay’s ability to rate buyers can allow service providers to mitigate their risk by adjusting their prices or terms of service to account for that risk. Service providers often do this in traditional transactions, where an electrician may charge a higher rate to a customer with a history of late payments. Applying this concept to automated service offerings may be a benefit to bother service providers and to higher-rated consumers.

3) Other: During our research, we have identified a capability gap that is, to some extent, a side effect of the fact that neither recreational social media nor existing business registries were designed as places for advertising and discovering service offerings. We believe there is a place for a business service registry that is analogous to the web service registries built on Universal Description, Discovery, and Integration (UDDI) standard. Such a “social UDDI” registry, built on a generic service description such as that we proposed in [5], would provide a mechanism for the discovery of services and recommendations based on users’ social and business connections.

V. UNTAPPED POTENTIAL

Our research indicates that there is untapped potential across the spectrum of social network applications and how they can be leveraged by both service providers and consumers to serve as matchmaking tools, both from the service discovery aspect and from the provider recommendation aspect.

A. Business Aspects of Social Media This paper is an outgrowth of our research into automating

workflow composition from available service offerings [4], [10]. We find that Facebook’s Graph API offers application developers a powerful interface into users’ social networks that can be leveraged to find potential service providers. However, this power is limited partly by the lack of service description support in Facebook and partly by the limited integration between a given business’s online offerings and their Facebook presence.

Our research indicates that business’ embrace of social media has only scratched the surface of the potential benefits. Putting up a simple Facebook page is the electronic equivalent of advertising in the local newspaper, albeit more interactive. Those businesses that have most aggressively embraced social media use their Facebook page as a means of communicating with their customer base by regularly posting updates regarding developments at their business, offering special deals to users who “Like” them on Facebook, or by soliciting users’ opinions using polls and other means of garnering feedback.

Even so, Facebook penetration into the business space remains relatively small. The barrier to entry is extremely low as compared to other mechanisms for reaching such a large audience, and very few advertising venues offer the kind of interactive experience that social media offers and that a younger generation of consumers has come to expect. We believe the main reason business penetration remains relatively low is that Facebook is designed primarily as a “recreational” social network and is not optimized for business use.

B. Social Aspects of Business Media Approaching the subject of untapped potential from the

perspective of the service provider, we find that while there is some penetration of service offerings into the recreational social media, there is very little service offering penetration into the business-oriented social media.

One example of this untapped potential is LinkedIn, the social media network aimed primarily at connections among business associates and colleagues. LinkedIn focuses on making personal connections among businesspeople but has very little provision for businesspeople to connect their businesses. Many businesses form teams with other firms to pursue work that neither business can adequately perform on its own; by the same token businesses often look for firms that can provide specific services that are outside their own expertise (such as a business looking for a payroll services firm).

LinkedIn allows a business to create a very simple presence on their network that is little more than a short description of

13131281

Page 5: [IEEE 2012 International Conference on Advances in Social Networks Analysis and Mining (ASONAM 2012) - Istanbul (2012.08.26-2012.08.29)] 2012 IEEE/ACM International Conference on Advances

their business. By expanding the level of detail available for business descriptions, LinkedIn could serve as a service registry where individual businesspeople or firms could discover services and providers. This may have an additional benefit in that it would allow users to recommend service provider. It currently allows a user to recommend another user, but there is no provision for recommending a business. This might have an additional benefit of improving business’ ability to attract talent; a business being recommended by businesspeople the user knows would indicate that may be a business the user would be interested in joining.

Likewise, eBay could increase its effort to function as a service broker; as discussed earlier the “services” section of eBay is small and not very well populated. As described above, eBay already has many of the aspects necessary to provider service offerings and service recommendations (as well as consumer recommendations to service providers). All that is missing is additional services available on eBay.

C. Social Services Broker As a potential remedy to these issues, we propose an

application based on Facebook Connect that can integrate services offered in a UDDI-type service registry with a user’s social network (including the businesses that user has a Facebook connection to). This application would allow users to specify workflow requirements, perhaps in a way similar to that described in [10]. Once the workflow requirements are posted, a services broker could leverage the Facebook Graph API to find service offerings within the user’s network. Upon matching services to the requirements specified in the workflow description, the application can then generate potential workflows by chaining the available services.

Enabling this sort of service selection and composition requires some machine-readable service description, analogous to the Web Services Description Language (WSDL) documents used by web services. Some services can be offered as either electronic of physical services; for example, a payment service may accept electronic funds transfers or it may require the user to send a paper check. To address this, we have developed a generic service description specification that expands on the concepts in OWL-S [11], [12] to describe services of all types. One view of the resulting service description can be seen in Fig. 3.

Fig. 3 General service description

Once potential workflows have been identified, the

application can then query the user’s network to generate recommendations based on the opinions of the user’s social network as well as more objective factors such as cost. These factors can be combined with user preferences to generate an optimized service recommendation that makes best use of the available services to generate a workflow. For example, a user may want to minimize cost, and be willing to combine any selection of services from any provider so long as the total cost is minimized. Alternatively, the user may the value convenience of dealing with the minimal number of service providers, and be willing to pay a premium to minimize the variety of providers in the resulting workflow.

This idea could be expanded to use an eBay-style auction where service providers can offer their services to the highest bidder. Another possibility would be to structure the bidding as a “Dutch auction,” where the service provider progressively lowers the price until a consumer is willing to pay that amount for the offered service. This latter option might be particularly attractive to web service providers. For web service providers the service offered is standardized and the marginal cost of offering the service to an additional consumer is minimal; a Dutch auction would help maximize revenue for a given service offering while rapidly adapting to changing market conditions.

VI. CONCLUSIONS The research we describe here is part of our work to

automate the composition of workflows from available services based on semantically-annotated descriptions of those services. As we describe here, the selection of available services extends beyond the realm of web-based services to include services that may be offered as physical interactions. We describe our investigation of the use of social media as a tool for both finding available services and for generating recommendations from among the available services.

We further describe possible applications that can improve the user’s ability to find and compose services, as well as improve service providers’ ability to improve visibility of their services to potential consumers. While privacy is a concern in all social media, the information necessary to develop direct recommendations (that is, recommendations based on the opinions of those in a user’s social network) presents no significant privacy threat. Expanding this idea to develop recommendations based on secondary sources (e.g., a service recommended by a friend of a friend) does present some privacy concerns based on the amount of information about each friend’s social network that would be revealed by granting such access.

VII. FUTURE WORK The research described here is part of an ongoing effort,

and we plan to implement some of the ideas described here into a prototype workflow composition system that uses the service descriptions shown here to find the workflows that can be constructed from available services, evaluate the

13141282

Page 6: [IEEE 2012 International Conference on Advances in Social Networks Analysis and Mining (ASONAM 2012) - Istanbul (2012.08.26-2012.08.29)] 2012 IEEE/ACM International Conference on Advances

alternatives based on the services’ characteristics as compared to user preferences, and recommend the optimal service composition based on those factors.

Based on the results described here, we think there is additional value that can be derived from the information available within users’ social networks, and that exploring the intersection of social media applications and service offerings may yield significant additional value for both service providers and service consumers.

REFERENCES [1] A. Tari, I. Elgedawy, and A. Dahmani, “A dual-layered model for web

services representation and composition,” Journal of Intelligent Information Systems, vol. 32, no. 3, pp. 237–265, Jun. 2008.

[2] R. Howard and L. Kerschberg, “A Framework for Dynamic Semantic Web Services Management,” International Journal of Cooperative Information Systems, Special Issue on Service Oriented Modeling, vol. 13, no. 4, pp. 441–485, 2004.

[3] S. Kona, A. Bansal, and G. Gupta, “Automatic Composition of Semantic Web Services,” in Web Services, 2007. ICWS 2007. IEEE International Conference on, 2007, pp. 150–158.

[4] J. McDowall and L. Kerschberg, “A multi-agent approach for generating ontologies and composing services into executable workflows,” in Proceedings of the 2010 EDBT/ICDT Workshops, Lausanne, Switzerland, 2010, pp. 1–12.

[5] J. McDowall, “More than IT: Extending SOA to the Entire Enterprise,” 3rd Annual DoD SOA & Semantic Technology Symposium, Springfield, VA, 13-Jul-2011.

[6] P. P. Maglio and J. Spohrer, “Fundamentals of service science,” J. of the Acad. Mark. Sci., vol. 36, no. 1, pp. 18–20, Jul. 2007.

[7] J. D. Montgomery, “Social Networks and Labor-Market Outcomes: Toward an Economic Analysis,” The American Economic Review, vol. 81, no. 5, pp. 1408–1418, Dec. 1991.

[8] R. J. White and A. E. Green, “Opening up or Closing down Opportunities?: The Role of Social Networks and Attachment to Place in Informing Young Peoples’ Attitudes and Access to Training and Employment,” Urban Stud, vol. 48, no. 1, pp. 41–60, Jan. 2011.

[9] A. Heß and N. Kushmerick, “Machine Learning for Annotating Semantic Web Services,” presented at the 2004 Spring Symposium, Menlo Park, California, 2004.

[10] J. McDowall and L. Kerschberg, “Agent Negotiation Strategies for Composing Service Workflows,” presented at the 28th IEEE International Conference on Data Engineering Workshop on Data-Driven Decision Support and Guidance Systems (DGSS), Washington, DC, USA, 2012.

[11] D. Martin, M. Paolucci, S. McIlraith, M. Burstein, D. McDermott, D. McGuinness, B. Parsia, T. Payne, M. Sabou, M. Solanki, N. Srinivasan, and K. Sycara, “Bringing Semantics to Web Services: The OWL-S Approach,” in Semantic Web Services and Web Process Composition, 2005, pp. 26–42.

[12] M. Paolucci, M. Wagner, and D. Martin, “Grounding OWL-S in SAWSDL,” in Service-Oriented Computing – ICSOC 2007, 2009, pp. 416–421.

13151283