Multi-Agent Systems (MAS) has grown into an interdisciplinary field that includes various tracks and embraces many previously disparate research areas. Particularly, multi-agent coordination, a sub-area of MAS, investigates how multiple intelligent computational agents work together to achieve high-level goals beyond the capabilities of single agents. Many different approaches have been investigated, such as the partially observable Markov decision process (POMDP), task structure analysis, coordination communication protocols, game theory, synergy, etc. Collaboration Technologies and Systems (CTS) have evolved significantly as well. These tend to investigate the design and development of effective environments or tools that help human users work together in a distributed collaborative, possibly virtual, fashion. Some notable examples of CTS include Collaboratories, collaborative design/editing, and on-line collaboration tools and environments. CTS is beginning to address the challenges of supporting coordinated, purposive activities. MAS is still facing challenges of scaling to large numbers of entities and real-world tasks (see, for example, James Hendler's inquiry, “where are all the intelligent agents?"1).
This workshop will explore potential overlap between CTS and MAS, and coordination because they share a common ground: how multiple entities - intelligent agents or humans alike - work together to carry out potentially related tasks. We will ask questions of whether and how to design and develop collaboration systems, promoting coordinated human activity, could be enhanced by incorporating insights from MAS. Collaboration technologies embody practical considerations from the human users' points of view, allowing users to ignore how the underlying (agent) infrastructure is implemented. Meanwhile, MAS/coordination investigates intelligent agents’ underlying algorithms and mechanisms and, in some cases, how artificial agents can interact with people as peers. Conversely, intelligent agents will not see significant acceptance, nor will they be able to manage the complexity and knowledge-intensity of meaningful practical applications, without developing some understanding of how to make effective use of human contributions throughout the specification, execution, evaluation, and refinement stages of the software lifecycle.
This workshop solicits papers that discuss issues underpinning overlaps between MAS and CTS, possible advantages/disadvantages of hybrids between them for designing and developing modern distributed collaborative software systems, and research and/or real-world experience and/or applications and/or lessons learned that involve both CTS and MAS. That is, any paper that addresses both CTS and MAS, preferably in one or a set of applications that share similar underlying research challenges, is of special interest to this workshop. An example could be: the design and development of a collaboration environment (i.e., a distributed planning tool) that enables multiple heterogeneous, human experts and agents to work in combination across complex computer networks on courses of actions in response to cyber attacks. Another example might be systems or interfaces supporting the division of labor between CTS and MAS elements during execution.