[Paper Review #9 7/12/16] Accessible Skimming: Faster Screen Reading of Web Pages

Accessible Skimming: Faster Screen Reading of Web Pages


Faisal Ahmed, Yevgen Borodin, Andrii Soviak, Muhammad Islam, I.V. Ramakrishnan, Terri Hedgpeth
UIST 2012, 11 pages except references


Summary

Sighted users use skimming technique to extract information from articles fast and decide to read further or not. It helps them to efficiently access necessary information in information-overloaded environment, e.g., web. However, blind users' reading experience is commonly limited to 1) either listen to all of the content, 2) listen to the first part of each sentence or paragraph before they skip to the next one. So that their skimming strategy is increasing the speech rate and using shortcuts to navigate by character, word, line, paragraph, section, page, etc.[1] There are two goals to use skimming technique: 1) looking for specific information, and 2) getting the content's gist. To help blind users achieve those goals, the authors suggest auto text summarization (skimming) algorithm called SkimSentence, and compare it to Gold/Human Skimming and Original paragraph.


Experimental Set Up


  • Web news articles. no link or image to reduce the variation among the articles. 
  • Same voice and speed through tasks. 
  • Participant: 23 ppl, varied ages, 3 different computer competence (experts to mildly comfortable)


  • Task 1: Listening and comprehension 
    • 1 original paragraph and 2 summaries
    • 12 comprehension questions: 1 topic (open-ended), 4 noun, 3 verb, 2 values, 2 adjectives/adverbs (multiple-choice). 


  • Task 2: Searching
    • 1 original paragraph and 2 summaries
    • Rephrased question is given before reading the articles - the answer is in last paragraph of the article. Measure time until finding the answer and keystrokes


Further to read


  • Borodin, Yevgen, et al. "Hearsay: a new generation context-driven multi-modal assistive web browser." Proceedings of the 19th international conference on World wide web. ACM, 2010.



Comments