Email with a mind on the future

27 Apr 2005

A University College Dublin (UCD) scientist has filed a patent application for a new technology that he believes can turn email into a much more effective business tool.

US-born Dr Nicholas Kushmerick, (pictured) a senior lecturer in the Department of Computer Science at UCD, has developed the technology over the past year during his part-time position as visiting scientist on IBM’s Centre for Advanced Studies (CAS) initiative. This is a programme that aims to forge links between the computer maker’s Dublin software lab and the academic community in order to turn new technology into marketable products.

Kushmerick developed the technology, known as Active Email Manager (AEM), in concert with a New York-based IBM researcher, Tessa Lau. Together they developed an algorithm that automatically tracks tasks and associated emails in order to build up a work-flow for each task.

“The vision is that rather than come in and download all your emails, you could just call up your to do list and manage your activities,” explains Kushmerick, who notes that while email is an excellent communications tool it has serious shortcomings as a teamworking platform.

At the same time, the computer program has the intelligence to distinguish between these work-related tasks and other types of email such as personal correspondence. Kushmerick stresses that these non-work emails are left intact and not deleted, read or otherwise interfered with by the software program.

Nine months into the project, Kushmerick and Lau have developed a prototype and conducted a series of experiments to measure the system’s accuracy and effectiveness. The two scientists have also presented their findings at the Intelligent User Interfaces conference in California in January — a leading international event for this area of research — where their paper was voted one of the two best papers at the event.

Kushmerick believes the technology could now develop in one of two ways: either as a fully automated tool for examining a user’s email inbox or as a semi-automated system where some user involvement would be required to refine and support decisions made by the algorithm. He picks the latter as the more likely route because the algorithm is not perfectly reliable or accurate and so probably would require some human intervention to become useful in practice.

IBM, meanwhile, has filed patent applications for AEM in a number of countries and beginning to explore its commercial possibilities. The technology is currently being appraised by two separate research groups within IBM. One of these is the Massachusetts-based product development team that develops IBM’s suite of collaboration software, Lotus Workplace. “There are some pretty intensive discussions going on now to see if we can get enough attention and convince them that our idea is feasible and that they would put it into their product pipeline,” notes Kushmerick.

The other research group showing interest in the technology has been working to develop a better user interface for email to allow office workers to deal with the growing email more effectively. The team sees AEM as an automated means of interpreting and categorising incoming mail, says Kushmerick. “The promise of our techniques is to automate large parts of that so that you can just show our algorithm all the messages that relate to asset management, for example, push a better and the system discovers that patterns, similarities and patterns and so on.”

While IBM is viewing AEM’s potential in the context of its own email client, Lotus Notes, Kushmerick also believes the software program has wider applicability. For this reason, he and Lau each plan to take on a research intern this summer to help develop a prototype of AEM that would work with an open source email client called Thunderbird. “This would make the technology available to many more people than if it was only done within Notes,” he explains.

Kushmerick has lived in Ireland since 1998. Having completed his PhD at the University of Washington in Seattle a year earlier, he was pondering his next career move along with his Canadian-born wife. They chose Ireland randomly, he says, by “throwing a dart on a map”. At the time Irish science was poorly funded but the advent of Science Foundation Ireland (SFI) has changed all that and made Ireland one of the best places in the world to do research believes Kushmerick, who doubts whether he would still be in Ireland were it not for SFI. “When I arrived in Dublin there was a modest amount of funding for computer science. With SFI it’s completely changed. SFI gives some of the biggest research grants anywhere in the world … the equivalent funding agencies in the US don’t give anything similar to the level of funding that’s available from SFI now. It’s really a fantastic opportunity for Ireland.”

As an SFI principal investigator, Kushmerick has received five-year funding (2002-2007) of more than €1m to build a team of nine scientists and conduct research in the area of machine learning and information extraction and retrieval. This focuses on the computational analysis of ‘natural text’ such as email messages and news articles.

“Natural text contains a lot of content but it’s all loosely structured,” he explains. “Computers understand structured data, where you have rows and columns of data with very well-defined meaning. Most of my work involves taking that unstructured data and somehow massaging into a structured format so that a machine can take some action and make decisions.”

By Brian Skelly