MediaGLOW: Organizing Photos in a Graph-based Workspace

Written by: Andreas Girgensohn, Frank Shipman, Lynn Wilcox, Thea Turner, Matthew Cooper

The authors designed a thing called MediaGLOW (Graph Layout Organization Workspace). This is an interactive workspace for sorting and viewing pictures. MediaGLOW allows people to view photos by similarity by letting the user group photos in a stack, then similar photos will group around the "stack." The distance that pictures group around the stack can be determined by either picture similarity or geographic distance. To normalize how far pictures are from the stack so that they dont tend to cluster at a single distance the authors find the percentage of similarity.



This research is important because it advances how people interact with media and are able to view and sort their large libraries of pictures. The downside is that this seems purely recreational and not useful to anyone that is working with pictures seriously. Future work should include synchronizing with online vendors like flikr or facebook.

Discovery-oriented Collaborative Filtering for Improving User Satisfaction

Written by: Yoshinori Hijikata, Takuya Shimizu, Shogo Nishida

The authors created a discovery oriented content filtering system so that the result is not information that the user already knows. It was found in the past that people stop using content filtering because it narrows the searches to very similar results and doesn't allow for much discovery. The author's research focused on prediction of unknown items, recommendation of items from the user’s preference and acquaintance, and examination of user satisfaction.

The two main content filtering algorithms are user-based and item-based. The authors used the prediction-combining algorithm and Independently Evaluating Algorithm (IEA) of different types to try and get a better response by joining the common results and filtering out what the user had probably seen before. The authors conducted an experiment using 20000 rating data collected from 100 users. They found that the IEA gave the best results and that combining their predictive algorithms with standard content filtering the quality and uniqueness of results goes up.

This is important because much of what people do on the internet is search for answers, hence google became popular. The problem with this research is that content filtering has never had a wide enough scope for an experienced searcher and this research didn't show very much difference between old CF results and their new one. I would think that you would have an algorithm that lets you vote things down as irrelevant and then sort out what content isnt good after it gets a good idea of what your looking for.

Timing Is Everything? The Effects of Timing and Placement of Online Privacy Indicators

Written by: Serge Egelman, Janice Tsai, Lorrie Faith Cranor, Alessandro Acquisti

The authors of this paper did research into how to effectively display privacy information for a web page so the user is best notified. Current privacy notifiers are just seals on a website. The tell almost nothing about the site's actual implementation of security practices; the seal just means there is a seal there. Studying how people reacted to these sites the authors found that most people base their idea of security from the "look and feel" of the website. The authors then constructed a study with four different groups. These were denoted by the security indicator being either just in the search results, in a small window after visiting a site as well as in the search results, as a full page pop-up after visiting a site as well as in the search results, and telling the user that the security indicator was actually a handicap accessibility indicator. The participants of the experiment then would search using a provided rigged search engine to purchase items. People only paid a premium for items from a secure page if there was a privacy indicator. Timing seemed to matter as people that saw the security indicator earlier were more likely to pay a premium, though people who searched multiple sites almost always paid the premium for security. People who saw a full page privacy warning usually paid a premium for all products that they bought. Also, people who had pop-ups of any nature to show privacy information visited 203% more websites to decide where to purchase from.



This paper is important because the every day user doesnt understand just how insecure many of the sites that they purchase from are. The downside to this research is that it only deals with getting people to recognize security as an issue, it doesnt construe any information about the actual security of the website. In the future I would provide a more in depth response available accessible by clicking on the security indicator and find out how many people just trust the image without checking any real settings.

From Geek to Sleek: Integrating Task Learning Tools to Support End Users in Real-World Applications

Written by: Aaron Spaulding, Jim Blythe, Will Haines, Melinda Gervasio

The authors created a integrated task learning system (ITL) that learns a task by watching the user. This is accomplished by using a macro reader to follow the actions, programming by demonstration (PBD), and a good UI so people will use it. The UI was key because people initially saw the ITL as "a bunch of geek stuff." Because people are unwilling to show a program how to do stuff more than once the ITL uses a very strong learning database for background knowledge and understanding.

This is extremely important because it can take care of menial computer tasks for companies and free up another minimum wage job. The program seems like it would be hard to use, similar to computer vision training where it takes a lot of work for a machine to know what it's looking for. Future work would be to expand on what functions are available to the users and refined learning capability for the ITL.

MusicSim: Integrating Audio Analysis and User Feedback in an Interactive Music Browsing UI

Written by: Ya-Xi Chen and Andreas Butz

The authors put together a program that sorts music into genres. It computes similarity between songs using jAudio, then it tries to sort "smart" by using Simple K-Means to place songs in a genre cluster. A user is allowed to type in a song name and it shows songs similar to that song in a cluster. They found that generally people liked the program and useful.

This paper is interesting because it expands on what pandora is doing. Maybe their algorithms will be good. The problem is that intelligently sorting music is nigh impossible because people's tastes are so peculiar so this project is doomed to be less than perfect. Future work will probably include machine learning but I feel they need to look at other ways of sorting the music than in clusters like they have chosen.

Emotional Design

Written by: Donald Norman

Donald starts the book off talking about how he was wrong about being able to just design things so that they function well. He found that even if they function fantastically, if they are ugly people will not use them because there is no aesthetic appeal and we are turned off to everyday use. He describes three ways to now look at design; visceral, behavioral, and reflective. Visceral being visual stimulation, behavioral being the liking of the actual interaction, and reflective being the logic behind why you like it. He then rants about this for 4 chapters with various examples. For the last couple chapters of the book he talks about the advent of AI and what it means to be alive and how I, Robot could come to happen in the near future. Woot.

This, again, is a design book, and we are told stuff that we know now because Apple has made such a large impact on everyday design. People in computer science right now read about how we need to design things and think about why things are marketable and why certain products fail. It is a good read if you don't already follow current computer trends, or the evolution of the computer world over the last decade.

Predictive Text Input in a Mobile Shopping Assistant: Methods and Interface Design

Written by: Petteri Nurmi, Andreas Forsblom, Patrik Flor´een, Peter Peltonen, Petri Saarikko

The authors put together a shopping list creator that specializes in predictive text and usability. They built a predictive text database and ran a usability test for both one and two handed use. The test showed significant speed improvement was achieved with both hands by using predictive text. Using predictive text improved the input error rate by 80%.

This paper shows that people have not put that much effort into grocery list building apps. I feel that this paper should have compared their database and correctness to other similar applications on the market. This also doesnt feel like a complete research topic to me. All the authors did was put a database together and rank items by frequency of use. Future work should include matching words to pictures and sorting lists by common food groupings.