Paper reading #24: Have a say over what you see: evaluating interactive compression techniques

Comments

 
Reference Information
Title: Have a say over what you see: evaluating interactive compression techniques
Authors: Simon Tucker, Steve Whittaker
Presentation Venue: IUI 2009: Proceedings of the 14th international conference on Intelligent user interfaces; February 8-11, 2009; Sanibel Island, Florida, USA

Summary

This paper discusses Interactive Compression techniques that allow a user to remove information in documents they do not wish to see and making important information more visible.

Some of the different compression techniques explored are Word Excision(removing unimportant words by replacing them with periods), Utterance Excision(removing utterances and replacing them with white space), Highlighting Words(marking important words), Highlighting Utterances( marking important utterances), Keyword Context(use Utterance Excision but leave one word in place and grayed-out to let the user know the utterance is there), Fish-eye View(split the screen into five views that display certain things such as important words).

After an initial user study, the authors found that two most successful compression techniques were Word Excision and Word Highlighting. They performed another more in-depth study on these two specific techniques.

Their results showed that Word Excision and Word Highlighting allowed the users to extract the important information from documents more effectively even with the presence of an occasional error from the algorithm

Discussion
While reading this paper, I could not help myself thinking about how useful a tool this software would be with all the HCI reading assignments we have had.
Areas of future study that the researchers mention include improving their algorithms and exploring other ways to determine which parts of the document are important. They also want to test their system on other types of documents.

Paper reading #23: Improving meeting summarization by focusing on user needs: a task-oriented evaluation

Comments

Reference Information
Title: Improving meeting summarization by focusing on user needs: a task-oriented evaluation
Authors: Pei-Yun Hsueh, Johanna D. Moore
Presentation Venue: IUI 2009: Proceedings of the 14th international conference on Intelligent user interfaces; February 8-11, 2009; Sanibel Island, Florida, USA

Summary

In this paper, the authors discuss a method of improving the summarization of meetings.

Much work on meeting browsers and search structures have been done that allow people to search for a specific part of the meeting, but often the most important part of the meeting is a decision that is made.  

The researchers discuss two types of summaries. The first provides a general summary of the meeting and is the type of summary produced by current. The second provides a more decision-focused summary that is shorter than the general summary.

The researchers performed a study in which they provided participants with four meetings through a Meeting Browser Interface. They asked the participants to summarize the decisions made in the meetings. Participants were randomly assigned one of four summary displays that were embedded into the browser interface that presented different information about the meeting.

The researchers found that displaying decision-focused summaries were more effective and helped users get a better overview of the meeting. However, it was found that decision-focused summaries written manually were still more effective than the ones generated by the algorithm.

Discussion
I think this research could be very useful for people who miss meetings or want to review meetings after they have occurred. Future studies suggests that the researchers will focus on further improving the algorithm and run more experiments on meetings that are more and less structured to better identify the strengths and weaknesses in their current algorithm.

Paper reading #22: User-oriented document summarization through vision-based eye-tracking

Comments

Reference Information

Reference Information
Title: User-oriented document summarization through vision-based eye-tracking
Authors: Songhua Xu, Hao Jiang, Francis C.M. Lau
Presentation Venue: IUI 2009: Proceedings of the 14th international conference on Intelligent user interfaces; February 8-11, 2009; Sanibel Island, Florida, USA

Summary

In this paper, the researchers seek to create an algorithm that allows eye-tracking to aid in summarizing documents for users. Their approach is by estimating the average time spent on a single word in the documents the user is reading and extrapolating that data into the likely-hood the user will find a sentence interesting. Equipment used consists in a web-cam, and the document viewer the authors made. After this setup, the user can start reading. They restrict the algorithm to only output a percentage of the sentences given the most attention. The percentage is based on the size of the document. 

The researchers compare their results of summarization to two popular methods of summarization - Microsoft Word AutoSummarize and the MEAD summerizer system.  The experiment involves using sets of literature from science and leisure

The researchers found that their algorithm could create more personalized summaries based on the user’s interests, especially for the entertainment/leisure articles, than the other two algorithms could.

 Discussion

This paper was really interesting to read. I appreciated the work presented, and the ideas put forward. They didn’t give much information on how the users responded to their system though. They just said it was better than the others.

The researchers note that in future studies they would like to make it so that their algorithm can work even on articles that the user has not read They also mention improving the overall algorithm and implementing a training system by allowing for feedback from the user to improve summarization.

 

Paper reading #21: Towards maximizing the accuracy of human-labeled sensor data

Comments

Reference InformationTitle: Towards maximizing the accuracy of human-labeled sensor data
Authors: Stephanie L. Rosenthal and Anind K. Dey
Presentation Venue: IUI 2010: Proceedings of the 15th international conference on Intelligent user interfaces; February 7-10, 2010; Hong Kong, China

Summary

In this paper, the authors discuss the impact that different amounts of information have on people when they label things. They discuss 5 main types of information given to labelers such a: Different amounts of contextual information, High and low level explanations, Prediction, User Feedback, and Level of uncertainty.

A study was conducted by the authors where they used the wizard-of-oz technique and presented labelers with varying amounts of information to test their labeling accuracy. They also focused on the differences between people labeling data they had not seen before and their own data.

After the study, they found that the five types of information had a positive affect on the labelers, because it gave them more information or helped to direct their thought processes.

Finally, the researchers found that whether the labeler was familiar with the had no impact on the accuracy of the label.

Discussion
The was one of those papers where I feel, if you do not have some type of background about the subject, you will be left in the dark. This paper was hard to follow, in my opinion. Nonetheless, the idea of labeling and the accuracy to which people do so was interesting. Either than that, I don't quite get the usefulness of this study.  I didn't understand why they were doing it as they did not have a good explanation in any part of the paper. 

Paper Reading #20: Designing a thesaurus-based comparison search interface for linked cultural heritage sources

Comments

Reference Information

Title: Designing a thesaurus-based comparison search interface for linked cultural heritage sources

Authors: Alia Amin, Michiel Hildebrand, Jacco van Ossenbruggen, Lynda Hardman

Presentation Venue: IUI '10 Proceedings of the 15th international conference on Intelligent user interfaces ;
February 7-10, 2009;

Hong Kong, China

Summary

 
This paper talks about comparison search, namely in the cultural heritage domain. The authors started with a preliminary study on experts trying to compare sets of artworks. From this study, the authors were able to identify certain areas of concern like name aliases, multiple languages, multiple terms, comparing many sets of data, single and multiple property comparison.  A thesaurus based algorithm solution is proposed in order to categorize data.  The uses are: learning about collections, planning exhibitions, and qualitative comparison.  They use a program called LISA in order to implement this. This paper mainly discusses the interface of the system and the interaction with it.
Some selection and comparison challenges mentioned are Searching artworks, Selecting artworks, and Comparing artworks. LISA was implemented on top of ClioPatria, a web-based application for searching through heterogeneous sets of data.

The researchers did two studies.  The first study was to get a grasp on problems experts have when doing comparison studies, and to determine the realistic use of comparisons.  The second study was to test a thesaurus-based comparison module that would be able to search, select, and compare different artifacts in cultural heritage.

LISA's ease of use was compared with RKDimages, a widely-used online cultural heritage archive. They found that overall LISA was easier to use than RKDimages, especially when it came to searching for many artworks and selecting artworksFuture works seem to indicate working on a better implementation of the auto-complete ability, supporting interactivity visualizations, and providing bookmarking and search history storage.

Discussion

 

I found this paper somewhat interesting. The target audience seems defined to art experts. Nonetheless,   the idea of using a Thesaurus to assist in search could be really useful. . It looked easy to use and looked self-explanatory.  One thing they could do for future projects would be to expand their domain to a different field.