Marco Mari

Enhancing information sharing through agents

This paper presents RAIS, a peer-to-peer multi-agent system for information sharing among a community of users connected through the Internet. RAIS offers a similar search power of Web search engines, but avoids the burden of publishing information on the Web and guarantees a controlled and dynamic access to information. The use of agent technologies has made straightforward the realization of three of the main features of the system: i) filtering of information coming from different users, on the basis of the previous experience of the local user, ii) pushing of new information that can be of interest for a user, and iii) delegation of access capabilities, on the basis of a reputation network, built by the agents of the system on the community of its users.

Collaborator - enabling enterprise collaboration through agents

This paper presents a particular aspect of the architecture of Collaborator, a software system that provides enterprise users with a shared workspace to support the activities of virtual teams. Collaborator exploits the seamless integration of everyday Web technologies with agents to promote flexibility in the interaction between users. In this paper we concentrate on the agent subsystem of Collaborator, and we show the architecture of this multiagent system and the functionality that it provides. In particular, we present the roles that agents play in the overall system (personal agents and session manager agent), and we detail on how personal agents are used to (i) manage the calendar of users; (ii) negotiate and classify meetings; and (iii) learn the preferences of users.

Agent-Based support for open communities

RAVE (Remote Assistance Virtual Environment) is a Web and multi-agent based system to support remote users during common projects or activities. A Personal Agent, associated with a specific user, helps her/him to solve problems proposing information and answers extracted from some information repositories. Its also able to forward answers received from other on-line users, recommended by their Personal Agents as experts in that specific topic. A profile is built and maintained for each user. A RAVE system is not a closed system, but its based on a dynamic network of RAVE platforms managing groups of geographically localized users and documents. RAVE users and documents profile management subsystems provide a mechanism that dynamically adapts the relevance of each profile, according to the availability of experts and documents.

Enhancing peer-to-peer applications with multi-agent systems

This paper copes with the problem of integrating peer-to-peer and multi-agent systems for the realization of both large-scale multi-agent systems and sophisticated peer-to-peer applications. In particular, it presents how JADE, one of the best known and most used software framework for the development of multi-agent systems, has been extended with a peer-to-peer technology, and how a JADE multi-agent system can be overlapped over a peer-to-peer system to provide more sophisticated services.

Enabling Vocal Interaction in a Web Portal Environment

The growing request of innovative, multimodal interfaces for mobile users and the need of different navigation paradigms for impaired Internet users are promoting and driving nowadays research on multimodal interactions. In this paper we present our experiences in the integration of vocal components in a portal engine, namely Apache Jetspeed. First, we discuss the reasons why the integration of a full-featured portal brings significant advantages to the development of a vocal platform. Then, we describe two complementary approaches for enhancing portals with vocal capabilities, and we compare them under various standpoints. The first approach that we investigate is based on a server-side speech synthesizer, while the second relies on client-side technologies, X+V and the SALT markup languages. For each of these approaches we present some examples and we discuss advantages and drawbacks.