Paper Reading #16: Mixture Model based Label Association Techniques for Web Accessibility

Reference Information:

Title: Mixture Model based Label Association Techniques for Web Accessibility

Authors: Muhammad Asiful Islam, Yevgen Borodin, I. V. Ramakrishnan

Conference: UIST '10 Proceedings of the 23nd annual ACM symposium on User interface software and technology

Summary:

The internet used to consist of text-based pages, but today the internet has grown into a massive chunk of data. The information is growing exponentially each day. However, many websites are not designed for blind users. An important aspect of making the Web accessible to blind

users is ensuring that all important web page elements such

as links, clickable buttons, and form fields have explicitly

assigned labels. Properly labeled content is then correctly

read out by screen readers, a dominant assistive technology

used by blind users.

Improperly labeled form

fields can critically impede online transactions such as

shopping, paying bills, etc. with screen readers. Very often

labels are not associated with form fields or are missing

altogether, making form filling a challenge for blind users.

Algorithms for associating a form element with one of several

candidate labels in its vicinity must cope with the variability

of the element’s features including label's location

relative to the element, distance to the element, etc. Probabilistic

models provide a natural machinery to reason with

such uncertainties. In this paper, the authors present a Finite Mixture

Model (FMM) formulation of the label association problem.

The variability of feature values are captured in the FMM

by a mixture of random variables that are drawn from parameterized

distributions. Then, the most likely label to be

paired with a form element is computed by maximizing the

log-likelihood of the feature data using the Expectation

Maximization algorithm.

The FMM approach was also adopted for two related problems: assigning labels from an external

Knowledge Base to form elements that have no candidate

labels in their vicinity and for quickly identifying clickable

elements such as add-to-cart, checkout, etc., used in online

transactions even when these elements do not have textual

captions e.g., image buttons without alternative text. The authors provide

a quantitative evaluation of their techniques, as well as

a user study with two blind subjects who used an aural web

browser implementing the approach.

The results of the experiments showed that after using VSM, 95% of the web elements were correctly labeled. Couple of blind users were asked questions and feedback to rate the system on the Likert scale. Both the users gave a very postive feedback and mentioned that they've always had problems filling out the online forms but this task was made easy after the web elements in the forms were properly labeled.

Discussion:

This idea is one example of how technology can be used to change lives. Technology has tremendous potential to benefit people. The paper does a great job at increasing the accessibility of the web for the blind users.

Overall, the paper was quite technical and presented the reader with complex mathematical formulas and code segments that were not quite comprehensible in the first reading.