TaskMinder
Team Members
Manas Tungare
Zachary Pousman [zpousman@cc]
Brian Landry [blandry@cc]
Rahul Nair
Proposal - Proposal Package
Final Project - Final Project Package
Work Area
Project Idea - The Task Mind
The Task mind attempts to create an order-of-magnitude leap forward in task management tools that exist today. The tools that exist today are suboptimal for a variety of reasons (including task entry, categorization, and in the ephemeral ?actually helpful in getting things done? way). But fundamentally, they fail because they don?t know very much at all about the user.
We seek to design an adaptive, personal ?Task Mind? that will apply machine-learning algorithms to the process of task suggestion and prioritization.
We plan to look at the following features of tasks to help create this priority-list of tasks and make smarter suggestions about what a user could do next:
- Time and Space data ? We suggest that certain tasks can only be done specific locations or at specific times, say, on Tuesday or only at the CoC. Our system will model user movements through space to learn where certain tasks are performed. We plan to track users on Georgia Tech?s Campus via software that locates users based on their use of the LAWN network.
- Start-time and End-time ? We seek to leverage statistics about when tasks were entered into the system and when they were marked by a user as ?completed.? This, we hope, can be used to tell how long tasks take (on average) and possibly group or rank the tasks by this criteria for suggestions based on time-available. This is hard because we don?t want to constantly bother users. How and even if this will be accomplished is still to be worked out. [other factors here?]
- Statistical properties of task entries ? The system will do some statistical analysis of the text of task entries to find patterns and similarities. We would use these, if they prove to be useful, in categorizing tasks into dynamically generated task domains (i.e., categories)
- Conversational Model of Interaction ? Though it is not entirely worked out how we'll do it, we feel that it would be best to provide feedback to the tool that is as rich as possible without making users crazy. We are thinking of a way to infer concepts like ?priority? / ?urgency? through a software interface that encourages the user to elegantly and easily tell the system what matters and what doesn?t.
Link to this Page