The Web for all Conference 2020 and the Web Conference 2020 - a retrospective

Jens Pelzetter's picture
Blogpost by: Jens Pelzetter

The Web Conference is one of most important conferences for research about the web. As part of this large conference the Web for All conference takes place, usually on the first two days. Along with the ASSETS conference, the Web for All conference is one of the most important conferences about research for accessible information technology.

The conference was planned to take place April 20 - 24 in Taipei (Taiwan), in the vicinity of the Taipei 101. Because of the COVID-19 pandemic the conference format was changed to an online-only conference. Thanks go to the organisers and volunteers for making this possible in a very short time.

Also because of COVID-19 all papers are freely accessible until the end of June:

Noticeable papers from the W4A

If you are creating videos for the web you should check out the paper Deaf and Hard-of-Hearing Users. Prioritization of Genres of Online Video Content Requiring Accurate Captions." by Larwan Berke, Matthew Seita, and Matt Huenerfauth. This paper presents the results of a study which investigated for which genres of videos deaf and hard of-hearing people deem captions necessary and for which genres they are not essential. The study used to following the list of genres: Animals and Pets, Autos and Vehicles, Comedy,  Education, Entertainment, Film and Animation, Games, How-To and Style, News and Politics, Non-Profits and Activism, Music, People and Blogs, Sports, Technology and Science, Travel and Events und Video-Blogging. The results are that for News and Politics, Education, Technology and Science, Film and Animation and Entertainment are necessary. For the genres Games, Animals and Pets, Sports and Music captions are not essential according to the participants of the study.

If you develop new skills for voice assistants you should check out the paper Accessible Conversational User Interfaces: Considerations for design by Kate Lister, Tim Coughlan, Francisco Iniesto, Nick Freear, and Peter Devine. This paper is a very good summary of the recommendations and best practices for Conversational Interfaces like chat bots or voice assistants.

As usual the Web Accessibility Challenge, in which novel solutions are presented, was quite interesting. The videos of the demos are available on Youtube.

For those who think that PDFs, especially if they contain mathematical formulas, are accessible: Please read the paper Towards Generating Web-accessible STEM Documents from PDF by Volker Sorge, Himanshu Garg, Neha Jadhav, Akashdeep Bansal, Ayushi Verma and M Balakrishnan. This paper discusses the accessibility problems of PDF documents and presents a system which converts PDFs into accessible HTML.

Two papers investigate the current tools for evaluating the accessibility of web pages. The first paper, Analysis of Automated Contrast Checking Tools by Rafael Almeida and Carlos Duarte compares different tools for checking the contrast on a web page. The second paper Comparing Accessibility Evaluation Plug-ins compares different browser extentions for checking web pages for accessibility problems. The recommendation from this paper: Use more than one tool. None of the tested tools covers all WCAG requirements.

Noticeable papers from the Web Conference

The Web Conference itself is a quite large conference with multiple parallel sessions. There I only saw a small part of the presentations. But here are some papers which I find very interessting:

  • Text-to-SQL Generation for Question Answering on Electronic Medical Records by Ping Wang (Virginia Tech), Tian Shi (Virginia Tech) and Chandan K. Reddy (Virginia Tech). This paper presents a deep learning model which can translate questions to SQL queries. Them model is specially tailored for the medical domain.
  • An Empirical Study of the Use of Integrity Verification Mechanisms for Web Subresources by Bertil Chapuis (UNIL-HEC Lausanne), Olamide Omolola   (TU Graz), Mauro Cherubini (UNIL-HEC Lausanne), Mathias Humbert  (armasuisse S+T) and Kévin Huguenin (UNIL-HEC Lausanne).   This paper presents the results of a study which investigated how many  web pages have adopted the Subresource Integrity Recommendation of the W3C.
  • Natural Key Discovery in Wikipedia Tables by Leon Bornemann (Hasso Plattner Institute),  Tobias Bleifuß (Hasso Plattner Institute),  Dmitri V. Kalashnikov (AT&T Labs - Research),  Felix Naumann (Hasso Plattner Institute)  and Divesh Srivastava (AT&T Labs-Research).  This paper describes a method for identifying the key column in wikipedia  tables. The context of table is also considered by the model.
  • Generating Representative Headlines for News Stories by  Xiaotao Gu (University of Illinois at Urbana-Champaign),  Yuning Mao (University of Illinois at Urbana-Champaign),  Jiawei Han (University of Illinois at Urbana-Champaign),  Jialu Liu (Google), You Wu (Google), Cong Yu (Google),  Daniel Finnie (Google), Hongkun Yu (Google), Jiaqi Zhai (Google)  and Nicholas Zukoski (Google).  This paper presents a method for automatically generating headlines for  news articles.
  • Abstractive Snippet Generation by  Wei-Fan Chen (Paderborn University), Shahbaz Syed (Leipzig University),  Benno Stein (Bauhaus-Universität Weimar),  Matthias Hagen (Martin-Luther-Universität Halle-Wittenberg)  and Martin Potthast (Leipzig University).  Because of some laws in the EU this paper is espcially interesting. It  describes a method for automatically generating summaries of other web pages.  The method does not extract the text, but generates a completly new text.
Tagging

Blog relation: