Machine access to audiology content on English Wikipedia: transcript

Key points

• Gabriel is a Brazilian audiologist studying for a master's degree at the University of São Paulo, working with colleagues from that university and from CDC NIOSH.
• Wikipedia is an open reference source that is also used to train large language models, retrieval-augmented generation systems and search engines.
• The aim was to map the availability and accessibility of hearing health information on Wikipedia that machines use to generate content.
• PetScan and the Pageviews tool were used to count articles and page views in categories such as sound, hearing, audiology, music, acoustics, ear and noise, at three levels of subcategories.
• More than 6,000 articles have some thematic relation to the categories, and the categories themselves total more than 900 articles.
• Sound, acoustics and hearing were the most viewed categories, by both humans and machines.

Chapters

1. 0:01Introduction and speaker background
2. 0:37Wikipedia as a source for people and machines
3. 1:14Objective and tools: PetScan and Pageviews
4. 2:59Categories analyzed and article counts
5. 4:26Page view results and conclusion

Full transcript

0:01Hello, everyone. My name is Hector Gabriel. I'm very happy for the opportunity to present this work at the VCCA this year. And in this season, I will present the abstract machine access to geologic content on English Wikipedia. I'm a Brazilian, as a UPL geologist, and I'm currently taking a master's degree at the University of São Paulo. and this study involves the participation of other resources from the University of Sao Paulo and from the CDC NIOSH.

0:37So, what is Wikipedia? Wikipedia is an open, collaborative, and general reference source on the internet. Available in multiple languages, it hosts millions of articles and receives billions of page views. It's important to note that Wikipedia serves as a training ground for large language models and also for retrieve argumentation generation systems and also as a reference for search engines on the internet.

1:14There is a history of activities and projects involving the use of Wikipedia and other Open knowledge projects maintained by the Wikimedia Foundation to promote hearing health. Wikipedia's usage is steadily increasing to include evidence-based information through educational activities and global runners' campaigns in the context of a hearing health content. In this sense, our objective was to map the availability and accessibility of human health information on Wikipedia, which is being used for machine-generated content. For this task, we use open-source tools such as pet-scan and page views to gather data from English articles categories on the Wikipedia related to audiology. This is a screen

2:16visualization of the Petiscan tool returning the number of pages on the English Wikipedia related to audiology category. The same process can be done for other categories for example articles linked to the listen or hearing or sound category of Wikipedia. This other screen shows, in the pageviews tool, the total of access, the total access metrics to the articles in the audiology category on the English Wikipedia in the last year.

2:59We use the PET scan and pageviews tools to analyze categories such as sound, hearing, audiology, music, acoustics, audiology, ear and noise. on English Wikipedia and that scan tool was used to list the articles with the demand categories two subcategories at three levels of content relationship div and page views supplied data on page views and two categories in two ways users so human visualization and machines searching genes and retrieved models and computer set programs such as bots The numerical values represent the quantity of pages on Wikipedia at each analyzed depth layer using the PadScan tool. More than 6,000 articles on the English Wikipedia have some distant thematic relation with the analyzed categories. And these categories together total more than 900 articles.

4:26In this other slide, the values represent the quantity of page views on Wikipedia for each category using page views tool. The most accessed categories were sound, acoustic and hearing. These are also the categories most accessed by users. As a conclusion, the access metrics showed and stood in the number of articles accessed by both human and machines justified activities aimed to promoting the creation of hearing health information on Wikipedia.

5:12I'll say that. Thank you very much for your attention and I'm available for any questions.

Source and licence

Recording: File:Machine Access To Audiology Content On English Wikipedia - 2024 Virtual Conference on Computational Audiology.webm by CorraleH (2024-06-21), via Wikimedia Commons. Licence: CC BY-SA 4.0. The transcript, summary and chapters on this page are shared under the same licence (CC BY-SA 4.0).

The transcript is machine speech-to-text by the TidyTools Audio & video transcriber; the summary, key points and chapters were written by Claude from that transcript. Both may contain errors: check the recording before quoting it.