Abstract for Michael Greenberg
|
|
|
Abstract for Michael Greenberg
|
Abstract For DTWA '97 Michael S. Greenberg Research and Advanced Development FTP Software, Inc.
Academic and industrial activity in the last few years has proven "software agents" can be sucessfully transported between computers. The next major challenge is to find useful applications for them. Although many people feel transportable agents will "revolutionize" the way we do business and interact on the global networks, this technological promise has yet to be realized.
Several issues must be addressed before any useful global deployment of transportable agent technology. Some of these include trust, reliable inter-agent communication and a recognizable context for global information. Of these, secure and trustworthy agent systems is perhaps the most important. Questions I have are:
o A common method of securing agent systems involves limiting what the agent can do, and therefore the potential harm from a "rouge" agent. How can an agent system be secure and still permit a range of agent activity to make the agent useful?
o The networking community has made the content of message communication very reliable. How can we make meaning and context of the communication reliable?
o The Internet is such a wealth of information and resources it is often hard for a human to navigate. How can this information be universally contextualized so agents can easily navigate it for their human masters.
I am interested in talking with other researchers about potential approaches for tackling these issues. Agents are currently "hot" in the public eye. If we do not deliver compelling agent-based applications soon we risk missing the window of opportunity.