Just sent my proposal to #NLnet's #ngi0commons to create a #Luanti #game to help with #Mozilla #CommonVoice clips validation.

The deadline is end of the current month. They should reply about the first round in the beginning August. And I should get a reply about the second round mid September.

It's about 300 hours of work, in the worst case. If things go smooth, I should have minimal working version for the White Cane Safety day which is on 15th October. And by end of the November, the first version is ready.

And all these are given that NLnet will give me the funding seeing my project valuable, and also not minding working with someone based in #Iran...

#FOSS #mozillacommonvoice #mcv #mozillacv #ai #ml #voicedata #crowdsourcing #crowdsourcingideas #minetest #minetestgame #luantigame #fossgaming #opensource #opensourceai #opensourcegame

For the past couple of years, as each new @mozilla #CommonVoice dataset of #voice #data is released, I've been using @observablehq to visualise the #metadata coverage across the 100+ languages in the dataset.

Version 17 was released yesterday (big ups to the team - EM Lewis-Jong, @jessie, Gina Moape, Dmitrij Feller) and there's some super interesting insights from the visualisation:

➡ Catalan (ca) now has more data in Common Voice than English (en) (!)

➡ The language with the highest average audio utterance duration at nearly 7 seconds is Icelandic (is). Perhaps Icelandic words are longer? I suspect so!

➡ Spanish (es), Bangla (Bengali) (bn), Mandarin Chinese (zh-CN) and Japanese (ja) all have a lot of recorded utterances that have not yet been validated. Albanian (sq) has the highest percentage of validated utterances, followed closely by Erzya / Arisa (myv).

➡ Votic (vot) has the highest percentage of invalidated utterances, but with 76% of utterances invalidated, I wonder if this language has been the target of deliberate invalidation activity (invalidating valid sentences, or recording sentences to be deliberately invalid) given the geopolitical instability in Russia currently.

See the visualisation here and let me know your thoughts below!

➡ https://observablehq.com/@kathyreid/mozilla-common-voice-v17-dataset-metadata-coverage

#linguistics #languages #data #VoiceAI #VoiceData #SpeechAI #SpeechData #DataViz

#PrivacyCon24 AI & Machine Learning panel was a little spicy...

Batul Yawer of ASU's work really stands out with research on the validity of a widely available #AI tool claiming "clinical grade performance" for stress and anxiety management.

Important questions as to deceptive marketing, health tool effectiveness, and potential harms as people rely on these tools to make health decisions.

https://www.ftc.gov/system/files/ftc_gov/pdf/15-Yawer-Reliability-and-Validity-of-a-Widely-Available-AI-Tool-for-Assessment-of-Stress-Based-on-Speech.pdf

#Privacy #HealthData #VoiceData #DigitalHarms #DeceptivePractices #FTC

📬 Active Listening: abgehörte Smartphones sollen Vorlieben der Nutzer preisgeben
#Datenschutz #ActiveListening #CMGLocalSolutions #CoxMediaGroup #VoiceData https://tarnkappe.info/artikel/it-sicherheit/datenschutz/active-listening-abgehoerte-smartphones-sollen-vorlieben-der-nutzer-preisgeben-284998.html

#Marketing Company Claims That It Actually Is #Listening to Your Phone and Smart Speakers to Target #Ads

“What would it mean for your business if you could target potential clients who are actively discussing their need for your services in their day-to-day conversations? No, it's not a #BlackMirror episode—it's #VoiceData, and #CMG has the capabilities to use it to your business advantage.”

https://www.404media.co/cmg-cox-media-actually-listening-to-phones-smartspeakers-for-ads-marketing/

This is a fascinating article on the increasing use of #subtitles, by Claudia Forsberg for ABC #Ballarat - the way that sound is designed for movies intended for cinema means that it doesn't play back optimally on mobile devices or streaming services - and this is one factor driving the adoption of #ClosedCaptioning or #Subtitles.

But, these #Subtitles are often inaccurate or mis-transcribed. They use #ASR technology - and this is another case for having good #voice #data #voicedata

https://www.abc.net.au/news/2023-02-12/subtitles-popular-among-general-population-change-tv-film/101956758

Voice Data is the Next Privacy Nightmare!

#News #TechNews #Privacy #biometrics #voicedata #cybersecurity

https://youtu.be/eO9nlnEl0jk

Daily podcast: Voice Data is the Next Privacy Nightmare!

#News #TechNews #Privacy #biometrics #voicedata #cybersecurity #podcast

https://soundcloud.com/nickaesp/vdp

Do you work with #voice and #voicedata, such as #data used for #ASR & #TTS tools? This might be collecting, cleaning, parsing data, or using data to train and evaluate #ML models 💻 🎤 📈

As part of my #PhD at ANU (not on fediverse), I'm doing some exploratory interviews to help shape a survey. The survey and interviews will help us understand #ML practice in voice, and create tools to help make voice tech fairer for everyone.

Ethics protocol: 2021/427

Would appreciate boosts / RTs for reach 💖