TaskTracer and Smart Desktop: Combining Activity Tracking with Machine Learning to Support Multi-tasking Information Workers

Speaker: Jon Herlocker , Oregon State University and Smart Desktop, Inc.

Date: Friday, September 21, 2007

Time: 2:00 PM to 3:00 PM

Refreshments: 3:45 PM

Public: Yes

Location: Patil/Kiva Seminar Room G449

Event Type:

Room Description:

Host: Rob Miller, MIT CSAIL

Contact: Michael Bernstein, x3-0452, msbernst@mit.edu

Relevant URL: http://eecs.oregonstate.edu/TaskTracer

Speaker URL: None

Speaker Photo:

Reminders to: seminars@csail.mit.edu, hci-seminar@csail.mit.edu, herlock@eecs.oregonstate.edu

Reminder Subject: TALK: TaskTracer and Smart Desktop: Combining Activity Tracking with Machine Learning to Support Multi-tasking Information Workers

A significant portion of today’s highly information worker’s life is spent in overhead: in particular trying to restore the context of a task, or re-find/re-access resources previously accessed. As the level of multitasking increases, information workers experience greater cognitive demand and greater overhead managing their activities and information across all their tasks. The TaskTracer research project at Oregon State University has been investigating the design of a software system to support multi-tasking information workers. The core vision of TaskTracer is to build software that understands that each user segments their life into tasks (contexts) that are personal to them; and knows, at every point in the interaction, which task the user is currently working on. TaskTracer combines tracking of interaction history with machine learning to adaptively learn models of the user’s tasks. TaskTracer’s user interface supports the user in getting to files faster, tracking time spent on tasks, organizing information, and recovering task context. The TaskTracer system is currently being commercialized under the name Smart Desktop. In this talk I will present an overview of TaskTracer research, discuss the challenges of integrating machine learning into the UI of an end-user product, talk about what I’ve learned while leading the product engineering team that is attempting to turn a research prototype into a commercial product, and do a short demonstration of Smart Desktop.

Speaker Biography:
Jon Herlocker is an Associate Professor of Computer Science at Oregon State University where his research interests revolve around helping people deal with overloads of information. Prior to TaskTracer, he was deeply involved in collaborative filtering research. His research has been funded by the NSF, Intel, DARPA, and other DoD agencies. He was a recipient of an NSF CAREER award in 2002. In addition to academic research, Herlocker is very interested in research commercialization and entrepreneurship. He has co-founded two startup companies – MusicStrands, Inc. (now MyStrands), and Smart Desktop, Inc. (now a division of Pi Corporation). He is currently taking a leave of absence from the University, and serving as Chief Technical Officer and Vice-President of Engineering of the Smart Desktop division of Pi Corporation.

Research Areas:

Impact Areas:

See other events that are part of the HCI Seminar Series Fall 2007.

Created by Linda L. Julien Email at Wednesday, June 19, 2013 at 6:22 AM.