Extra Credit Paper #1 (CHI 2010 #9)

Title: There's Methodology in the Madness: Toward Critical HCI Ethnography
Authors: Amanda Williams, Lilly Irani
Venue: CHI 2010, Atlanta, Georgia

Summary
This paper details some of the problems found in qualitative data collection from current ethnographies, and suggests more generalized approaches than those currently used. The complication in rhetoric of the user and the expansion of field sites that widen the scope of HCI and the issues these topics raise are also discussed as new developments within HCI.

The authors also discuss how the boundaries of field sites have changed, and ethnographers have taken on a more direct invasive approach recently. To address this, they propose that the criteria for ethnographies be revaluated. Below is one of the models for ethnographic research discussed in the paper:

Discussion
While I wasn't terribly interested in the topic itself, I think the author's work is imporant. Certainly any time you feel that the validity of research in your field is being compromised you should speak out and do your best to address it. This paper might have been more interesting to read had I done so before our ethnography this semester.

Paper Reading #25: Using language complexity to measure cognitive load for adaptive interaction design

Title: Using language complexity to measure cognitive load for adaptive interaction design

Authors: M. Asif Khawaja, Fang Chen, and Nadine Marcus.

Venue: IUI '10 Proceedings of the 15th international conference on Intelligent user interfaces

Comments

http://isthishci.blogspot.com/2011/04/paper-reading-25-using-language.html
http://vincehci.blogspot.com/2011/04/paper-reading-25-email-overload.html

Summary
This paper was based heavily on the concept of cognitive load, the idea that a given task imposes a mental load on the individual working on it similar to the load on a CPU. The authors present an adaptive interaction system which measures this load, and adapts itself to improve user experience and performance. Cognitive load is measured through speech content analysis based on their language and dialog complexity. Because the medium used for measuring cognitive load was speech, operators (such as support) were analyzed for the data presented. A breakdown of data collected is included in the table below.

Discussion
I found that this paper was less about an adaptive system and more study focused than the authors originally claimed, which suggests that the abstract is somewhat misleading. As a study however, I feel their work is promising. There is obviously a correlation between speech and cognitive load, although more work should be done to investigate other ways of measuring cognitive load so that the results can be verified more strongly.

Paper Reading #24: Mobia Modeler: easing the creation process of mobile applications for non-technical users

Title: Mobia Modeler: easing the creation process of mobile applications for non-technical users

Authors: Florence Balagtas-Fernandez, Max Tafelmayer, and Heinrich Hussmann.

Venue: IUI '10 Proceedings of the 15th international conference on Intelligent user interfaces

Comments

http://isthishci.blogspot.com/2011/04/paper-reading-24-natural-language.html
http://vincehci.blogspot.com/2011/04/paper-reading-24-finding-your-way.html

Summary
This paper details a development tool termed Mobia Modeler designed to enable users without programing experience to create their own mobile applications. The tool features configurable components that are common in mobile applications. To illustrate the concept, this paper focused on health monitoring applications for mobile devices. The authors go on to describe a sample application called Health Monitor created using Mobia Modeler, and the steps taken to create the application. A labeled screenshot of the interface in use is shown below:

Discussion
I thought this was one of the most applicable projects of all the papers I have read in the intelligent user interfaces venue. While the quality of apps created by this program will almost certainly be lower than applications created by someone with programming knowledge, this type of innovation has great potential to take off. Blogs like this one are a similar concept, where most individuals don't have the time or experience to make quality websites, blogs provide a compromise between quality and features, and ease of use.

Paper Reading #23: Evaluating the design of inclusive interfaces by simulation

Title: Evaluating the design of inclusive interfaces by simulation

Authors: Pradipta Biswas, and Peter Robinson.

Venue: IUI '10 Proceedings of the 15th international conference on Intelligent user interfaces

Comments

http://detentionblockaa32.blogspot.com/2011/04/paper-reading-23-natural-language.html
http://vincehci.blogspot.com/2011/04/paper-reading-23-automatic-warning-cues.html

Summary
This paper details a proposed simulator for the design of assistive interfaces. The simulator can help predict possible interaction patterns when undertaking a task using different input devices in the presence of extraneous circumstances, such as disabilities. While the exact method of predication is not discussed in the paper, the authors describe the results of an experiment conducted with 7 individuals. Elements from figures 1 and 2 were isolated and displayed to each individual, and then they were asked to click on the same icon when it was displayed in a group.

The average relative error in response time was found to be 16% with a standard deviation of 54%. In 10% of the trials the relative error was more than 100%. Removing these outliers resulted in a average relative error of 6% with a standard deviation of 42%.

Discussion
I didn't like this paper very much. The general concept is interesting and somewhat unprecedented, as the authors are essentially attempting to quantify interface design, but the results were very poor in my opinion. Additionally, the sample size was too small to draw any meaningful conclusions from, and the paper states that all participants were trained for the experiment. Additionally, the authors appeared to think it was acceptable to remove 10% of the data points constituting outliers from the data analysis. If their system was implemented in interface design, I think a 10% chance for the system to just fail altogether on a prediction is extremely high, and not something that can be discounted.

Paper Reading #22: From documents to tasks: deriving user tasks from document usage patterns

Title: From documents to tasks: deriving user tasks from document usage patterns

Authors: Oliver Brdiczka

Venue: IUI '10 Proceedings of the 15th international conference on Intelligent user interfaces

Comments

http://detentionblockaa32.blogspot.com/2011/04/paper-reading-22-personalized-user.html
http://vincehci.blogspot.com/2011/04/paper-reading-22-pomdp-approach.html

Summary
The focus of this paper was on increasing the efficiency of multitasking in the average workplace. Existing systems for task management require a large amount of time investment from users to be effective, because the systems need to be 'trained' in order to work correctly. This paper proposes a new approach for automatically estimating a user's tasks from document interactions, without requiring access to the content of those documents. The system described in the paper instead looks at only the switches between which document is being interacted with by monitoring each user's activities and logging which documents had focus and when. Comparing this data allows the authors to build a similarity matrix based on document focus frequencies, dwell times, and switches. A clustering algorithm is then used to group documents into tasks based on the similarity matrix.

To evaluate the performance of the system, three values were used, called prevision, recall and F-measure. Precision refers to the faction of documents in a cluster that belong to the task label of that cluster, recall represents the fraction of all document that belong to a task label and appear in the corresponding cluster, and F-measure is the weighted mean of precision and recall. These values are graphed to the left.

Discussion
The system described in this paper seems promising. The authors described existing systems as either deficient, or requiring access to the content and title of each document in order to derive similarities from them, which raises security concerns in some businesses. It would be interesting to see how this approach compares to other systems that do analyze the content of each document in order to determine if the patterns associated with document access is sufficient to build similarities, or if more information is needed.

Paper Reading #21: iSlideShow: a content-aware slideshow system

Title: iSlideShow: a content-aware slideshow system

Authors: Jiajian Chen, Jun Xiao and Yuli Gao.

Venue: IUI '10 Proceedings of the 15th international conference on Intelligent user interfaces

Comments

http://shennessy11.blogspot.com/2011/04/paper-reading-21.html
http://detentionblockaa32.blogspot.com/2011/04/paper-reading-21-social-signal.html

Summary
iSlideShow is a system for automatically analyzing information from photo collections to create compositions and transitions similar to traditional slideshows. The authors describe two modes: a story-telling mode and a person-highlighting mode. The former mode instructs the system to use a theme-based clustering algorithm where multiple photos within a theme are tiled on the same slide. Different layouts are used and the slideshow is animated by transitions between clusters.

The second mode type, the system uses face detection algorithms to highlight faces within the photos and creates clusters based on the presence of certain individuals. Transitions in this mode are preformed between clusters of individuals within the photos.

To the left is an example interface for the slideshow. Each cluster is displayed as a sort of collage, and the location within the slideshow is displayed by the preview bar.

Discussion
While interesting, I felt this system had little to do with computer human interaction, or really anything beyond screen savers. The system described is obviously not intended for the creation of a presentation, and the uses of computer generated collages is fairly low. I didn't feel like the system really did innovated on anything either, as face detection algorithms are widely available through the Intel computer vision library (which is free to use). Themes were clustered by simply using time stamps on the photos, so besides utilizing well established algorithms for facial recognition, the system really isn't intelligent at all as the authors claim.

Paper Reading #20: Raconteur: from intent to stories

Title: Raconteur: from intent to stories
Authors: Chi, Pei-Yu and Lieberman, Henry
Venue: IUI '10 Proceedings of the 15th international conference on Intelligent user interfaces

Comments
http://angel-at-chi.blogspot.com/2011/04/paper-reading-20-rush-repeated.html
http://shennessy11.blogspot.com/2011/04/paper-reading-20.html

Summary
This paper presents a story editing system to help with video and picture editing for novices. The software, called Raconteur, helps users assemble coherent stories for media elements. Commonsense, a reasoning technique and database, is used to identify similar story elements and link them together. In other words, the interface provided is intelligent and attempts to automatically relate uploaded pictures based on a short description provided by the user. The picture below shows the interface of Raconteur.













Discussion
The interface presented in the article reminded me mostly of facebook, and I have a feeling that this is the audience the authors had in mind for the product. Unlike most articles, this one has the feel of also being a product in addition to being research, which makes me doubt the objectivity of the content. In any case, I like the general idea, even if the technology could be better applied else where. When it comes down to it, any system like this will get things wrong from time to time, and when it does users will spend much more time than they ordinary would have to fix it, or else wont notice the mistake at all. Intelligent systems are better used for data mining, such as Watson.