Human-Provided Services

Last update: Jun. 30, 2011

 

Keywords: Service-oriented computing, human-provided services, crowdsourcing, social networks, expertise mining

Introduction

The Web is evolving rapidly by allowing people to publish information and services. At the heart of this trend, interactions become increasingly complex and dynamic spanning both humans and software services. Thus, there has been a growing interest in the complex structure and dynamics of todays society. Our online-society is increasingly influenced by networks, incentives, and the behavior of social communities. However, the transformation of how people collaborate and interact on the Web has been poorly leveraged in existing service-oriented architectures (SOA). In SOA, compositions are based on Web services following the loose coupling and dynamic discovery paradigm. We argue that people should be able to define interaction interfaces (services) following the same principles to avoid the need for parallel systems of software-based services (SBS) and Human-Provided Services (HPS). We define such systems as mixed service-oriented systems.

The fundamental questions addressed by the HPS approach are:

  • Which of users activities are most relevant for provisioning human expertise?
  • How can people provide services in a personalized manner?

The benefit of this approach is a seamless service-oriented infrastructure of human- and software-based services. In this research, we focus on innovative applications based on mixed service-oriented systems. Specifically, we focus on service-oriented crowdsourcing in open enterprise environments. Dynamically changing properties including user preferences, changing expertise, and reputation make the design of mixed systems challenging. The novelty of our approach is that context-sensitive interaction mining algorithms track these properties based on monitoring of ad-hoc interactions.

Recent directions are formation principles and compositions in mixed systems. An interesting idea is the concept of structural holes that was introduced in a sociologicol context. Our proposed approach attempts to discover brokers in collaborative networks. The paper (see Bridging Socially-Enhanced Virtual Communities) introduces the broker concept in the context of mixed systems and a query language called BQDL (Broker Discovery and Query Language).

If you are interested in these topics contact Daniel Schall. Below you find some publications related to these topics.

Selected Publications