Chris Kam
Aaron Kirkes
Reference Information
Title: The Why UI: Using Goal Networks to Improve User Interfaces
Authors: Dustin A. Smith, Henry Lieberman
When/Where: IUI 2010, Hong Kong, China
Summary
This paper discussed modeling user goals to improve user interfaces. When an application is aware of the user's goal, the knowledge can be used for automated planning to help the user by automating common tasks. The idea is to construct a goal network, relating what goals people have with how people solve them. Goals are labeled with details such as where the goal is met and how long it takes to achieve. Goal networks have previously been used for recommendation systems, processing natural language queries, and for plan recognition systems.
Goal libraries are used to map interface-level goals to human-level goals (goals that people communicate with each other). This paper discussed processing natural language plans from a website where users share and discuss goals and plans in order to acquire data for the goal library. This is done by considering goal statements that specify a specific goal and plan statements, stories by people of how they achieved that goal. The English statements were corrected and tokenized, a directed, weighted goal graph was created then clustered, and verb-phrases were extracted and associated with rules. The human goals were analyzed to find which were most popular, controversial, and which goals were most central to the entire goal network. Finally, using this information, an example interface was created using to-do lists for plan recognition.
 |
| A subgraph for a single goal. Source: Article. |
This paper discussed more of an idea than an actual application, but it was still fairly interesting. I agree that goal-oriented applications could be very helpful to users, so focusing on collecting and parsing goals, then applying them to user interfaces seems like it could be a very useful application. The paper could have been more informative if it had focused more on application than explaining the subject area. In addition, the one application that was discussed was not described very well, so I am still a little unsure about how it works. Maybe including a brief user study could have helped understanding the topic or even have been useful to understand how important the subject area actually is. However, goal networking could be very useful for applications such as automating simple tasks and reminding users of steps that they have forgotten within a process, so I think that it should be kept in mind when designing user interfaces.
I think that the user's goal is something that should definitely be taken into account in interaction design, but it seems to me like dynamically determining user's roles and changing the UX accordingly seems like more trouble than it would be worth. Like you said though, we would need to see some actual implementations to know for sure.
ReplyDeleteI like that the paper discussed an actually application as well. Some of them talk to much theory
ReplyDeleteMy paper for this week was somewhat related to this. Researchers did a study about cognitive load as a basis for a future implementation of a system that would modify the user interface and interaction in order to make it easier for the user. Overall, I think the concept of modifying the user interface in any sense so that it will better fit the user is a great idea. However I'm not sure how efficient it would be. I guess it would depend in the way they develop the changes.
ReplyDeleteI'm not sure I see the point of designing an application that dynamically improves an interface. Just design a better interface. I feel that if this were to work, it would have to have a very application-specific implementation. You can't really quantify discrete goals when using a word processor, for example. Thoughts?
ReplyDeleteActually, you can quantify discrete goals, and they have, and then they made wizards. Sure, this paper proposes a level of abstraction away from having to explicitly click through a dialog, but again, the dialog is intuitive enough, and I seriously doubt that this type of application will reach that same level of intuition.
ReplyDeleteI think it was an interesting concept to explore. Though I believe if we could recognize a user's goal we could do even more than change the UI to help them along.
ReplyDelete