Enhancing e-Commerce with Intelligent Agents in Collaborative e-Communities

Abstract

The goal of this project is to design an intelligent agent-based framework for collaborative e-commerce applications that will better enable online shoppers to emulate the social experience of shopping in real space.

Research will be directed towards the development of a Multi-Agent System (MAS) architecture in which a number of geographically dispersed users, including both customers and merchants, can participate. Given the importance of collaboration and community in e-commerce, the architecture to be developed will be founded on both sound technologies and a carefully considered legal infrastructure.

Background

Considering the current growth of e-commerce on the Web and the desire to make shopping as easy, natural and enjoyable as possible, it would be interesting to enhance the way people currently shop on the Web by adding support for more collaboration between customers and salespersons or among customers. Providing an e-community web shopping experience would make on-line shopping closer to the actual experience people have in real shopping environments, one that they are likely to engage in more often and enjoy more.

One of the advantages of applying e-communities in e-commerce applications is the enhanced interactivity between merchants and buyers, and between customers and visitors. It would enable online merchants to offer features that are lacking in most of today's e-commerce stores. The community online shopping mall would make it easy for storeowners to provide real-time customer support, sales assistance, cross selling, promotion and individualized care, all services that have traditionally been proven to improve sales.

In addition, applying an e-community in an e-commerce application offers an opportunity to model intelligent agent behaviour along existing ethical and legal norms, a form of representation not currently applied in most existing e-commerce agent systems.

Project Description

The objective of this project is develop a multi-agent architecture for collaborative e-commerce applications; one that will better enable online shoppers to emulate the social experience of shopping offline and that respects ethical and legal notions such as informed consent and the protection of personal privacy.

As part of this project we will use intelligent agent technologies to develop collaborative e-communities; investigate and develop voice-enabled intelligent agents with emotive behavior; develop multi user-interface agents based on user models; conduct research on negotiation protocols for intelligent, autonomous agents; investigate protocols for multi-agent interaction in the context of e-commerce applications; and finally investigate, develop and apply ethical and legal principles to ensure that agent behavior and protocols comport with national and international consumer protection, as well as privacy and data protection legislations, respectively.

This project builds on earlier research that we have undertaken and is directed towards the development of a working prototype.

Project Significance

The research will have significant impact on the practical applications of intelligent-agent-based e-communities of buyers and vendors in the industry. It will also have a significant impact on the broad social norms underlying the design of agent systems.

Many current e-commerce applications exploit agent technologies to misrepresent online products and services or to surreptitiously gather and mine personal data. By focusing on consumer protection and privacy principles as a significant design feature, our research will re-orient the way that responsible technology creators conceive of the e-communities that they are not only constructing but also populating.

Approach and Methodology

In previous project work, an e-commerce prototype with an architecture for agent-based collaborative and interactive environments was developed and implemented. This project intends to continue and expand that research towards a working prototype.

To create a working prototype of e-commerce agents working in a community, we will first focus on a prototype for setting up agents and interactions between agents in the community, and then on a prototype with social and legal aspects added.

We will study user interaction models, and design a collaborative and interactive environment for information exchange among agents, including the design of appropriate agent interfaces.

The prototype system developed will use .NET technology. Agents will maintain a C# remote object proxy that will be used for all inter-agent communication, and a naming service will be developed to manage references to agents entering and leaving the system. Agent logic will also be built into the agents so negotiation protocols can be handled. User applications will also be developed as .NET web applications so that they can be run directly from most web browsers and interact easily with the agents in the system. The second prototype will build on the basic agent structure protoype developed earlier. It will focus on expanding the social and legal interaction between agents through the design of interaction and communication protocols for collaboration, negotiation, and profiling, and through the design of voice enabled intelligent agents. As collaboration among an organization of agents requires a common medium that is flexible, dynamic, and scalable enough to enable open communication, the second prototype shall provide multiple mediator agents, implemented as federated objects over the network, which will collaborate to service requests from applications.

As one of the goals of this project is to blend the technology and the law to fuse software code with legal codes, throughout the project, the technological team will work with the law and policy team to build privacy and consumer protection principles considerations directly into the system architecture. Fair information practices and other consumer protection principles will be built into the engagement protocols of the agents interacting within the e-community.

Finally, once ethical and legal considerations have been addressed, we will implement a working prototype, with architecturally key features, for initial local performance evaluation and refinement.

Research Team

The Research team assembled for this project includes Dr. Nicolas D. Georganas, Dr. Shervin Shirmohammadi and Dr. Ian R. Kerr, all of who bring unique and complementary expertise to the project The team also includes graduate student Dr. Xiaojun Shen, a research engineer; master’s students Xin Wang and Christopher Desmarais, SITE, University of Ottawa and Alex Cameron, Common Law as well as Co-op students, Xingyuan Wang and Zhecheng Jin.

• Dr. Nicolas D. Georganas, Distinguished University Professor, SITE, University of Ottawa, brings to the team expertise in Multimedia Applications, Multimedia Protocols, Intelligent and Autonomous Agents and Performance Evaluation. He will be responsible for overall research organization and project management, and also for research in interface design and intelligent agents.

• Dr. Shervin Shirmohammadi, is an Assistant Professor at SITE, University of Ottawa. He brings to the team expertise in Multimedia Communications, Software Engineering, and Distributed Computing and Communications. He will be responsible for detailed research methodology and supervising the software development process, including analysis and problem solving, work breakdown, implementation, testing, and integration.

• Dr. Ian R. Kerr, Faculty of Law, University of Ottawa and a member of the Law Society of Upper Canada, brings to the team expertise in Ecommerce and Internet Law, Privacy Law, Philosophy and Applied Ethics. He will be responsible for supervising the legal and ethical components of the project and for research on infusing norms into agent systems.