Thursday, February 25, 2010

CRAFTing an Environment for Collaborative Reasoning

(comment left on Patrick Webster's blog)

Susanne C. Hupfer IBM T.J. Watson Research Center, Cambridge, MA, USA
Steven I. Ross IBM T.J. Watson Research Center, Cambridge, MA, USA
Jamie C. Rasmussen IBM T.J. Watson Research Center, Cambridge, MA, USA
James E. Christensen IBM T.J. Watson Research Center, Cambridge, MA, USA
Stephen E. Levy IBM T.J. Watson Research Center, Cambridge, MA, USA
Daniel M. Gruen IBM T.J. Watson Research Center, Cambridge, MA, USA
John F. Patterson IBM T.J. Watson Research Center, Cambridge, MA, USA

Paper Link:

This paper covers research on collaborative reasoning and sense making on large scale "wicked" problems.
They described sensemaking as "a motivated continuous effort to understand connections (which can be among, people, places, and events) in order to anticipate their trajectories and act effectively."

The goal of their research on collaborative reasoning and problem solving led them to focus on the aspects of semantics, collaboration, and adaptability and designed a system to guide collaborative problem solving based on these goals.

To this end, the IBM team developed CRAFT (The Collaborative Reasoning and Analysis Framework and Toolkit) which allows for a generalized and visualized way to create an ontological model (basically an object oriented visualization system) that tracks relationships between entities. This visualization system provides a lengua franca (common tongue) to exchange information between members of an investigative team regarding a problem being solved.

Alongside entity tracking, the system can continually update and evolve existing data and meta data on the objects with continual inquiring and searching.

The system also allows awareness of multiple entities and users who are using the CRAFT system and see what inquiries and updates they have made to the system.

Making an inquiry into the system can uncover previously made inquiries by other entities and new entities that share a name can be flagged for either identifying the entities as the same entity or disambiguating the same-named entities.

CRAFT also provides investigation nodes that allow the user to question, hypothesize, inquiry, and gather evidence for a particular model, question, or investigation. These investigation nodes allow the user to model a particular scenario and gather evidence for or against that scenario.
(For example an investigation node on a corporation might include stock quotes relevant information and expert opinion on the action of the stocks.)


My spill:

This sounds like an interesting problem. You always hear in the news about multiple organization that are unable to cooperate and collaborate critical information leading to some disaster or another. The CRAFT system seems like it could solve these kinds of problems.

They did mention that they needed to implement a system that had access to the internet. I think that kind of feature is critical for a system of this kind just so the system can have access to the gigantic amounts of data available on the web and the ability to collaborate with multiple entities over several networks.

I could see a big problem with information security on a collaborative system especially for criminal investigations and the like.

No comments:

Post a Comment