I am a part of this year’s Oculus Launch Pad, a program for diversity and inclusion to help launch new content to market in VR on the Rift store. To my dismay, in the process of looking at my application submission (deadline in a few weeks so in super crunch mode), and in my discussions on calls with other staff, I am not happy about the data privacy policy.
" Data Collection
If your app collects data, you must clearly articulate your data privacy policy, and then abide by that policy. Comply with all relevant privacy laws. If you collect a user’s email address, the user must explicitly opt-in for both its collection and use. We are protective of our community’s email experience and do not want Oculus to be associated with unexpected email deliveries, so err on the side of user awareness. Of course, your app must abide by the laws of your country and the territories into which it will be published.
For mobile apps, be extra careful to remove any unused Android permissions from your application manifest, particularly those that relate to the user’s private data.
Note : We regret that Oculus cannot provide templates, legal advice, or other advice to assist in generating your privacy policy."
It looks like I need to write something of my own if I am collecting user generated data to help standardize Tagalog and pre-Philippine sanskrit (initial letters/characters to words, and eventually a word embedding as no corpus sort of exists with the full language, not even Google or Facebook to my knowledge has this sort of information as it’s never been statistically significant before for an endangered language writing system).
Any thoughts on
-
Accessible data privacy language - something easier to read than half the stuff I’ve read about GDPR in virtually every email list-serve that I have
-
This project is meant to generate the first 3D non-English dataset. 3DMNIST is the next closest thing. While I’ve read tons of papers on the subject for MNIST and many other 2D flat language datasets for other Asian languages, all of which, are much more data rich, I’m finding it difficult to find examples of other languages using ML and user-generated data from scratch to be able to assist me in the first step with NLP. Translation is not the same here (not all languages are created equal), and I’m wondering if there are any other folks working on similar projects even if not in 3D related to less widely known languages using NLP techniques to help with language preservation.
I previously spoke about VR Ethics at VRDC (VR Developer Conference) a few years ago and that is more along the lines of other forms of data that many startups are collecting from Virtual Reality (VR), that does include this sort of thing, but more along the lines of HIPPA compliance, eyetracking etc. My app is attempting to capturing user movements and strokes in VR of the actual calligraphy itself so that it does a sort of auto-draw to recognize the characters and eventually help create more words through our education app + creative experience (it’s like Tiltbrush/Medium/Quill from scratch plus an interactive storytelling piece) that eventually I think can form the beginnings of a corpus with other datasets for translation.
Any help or advice for data privacy language is much appreciated. Thanks!