ilyausorov

7 ms·

1.
by ilyausorov 2mo ago
Really awesome work! I've been trying to do some of this real time back and forth voice coaching myself and it's no easy feat. Congrats on the progress.
2.
BoldVoice (YC S21) Is Hiring Fullstack and Machine Learning Engineers(boldvoice.notion.site)
1 points by ilyausorov 8mo ago
3.
by ilyausorov 11mo ago
For sure the voice standardization model is not perfect, but it was important for us to do especially for the voice privacy. It’s still pretty early tech.
4.
by ilyausorov 11mo ago
Thanks, we love you too
5.
by ilyausorov 11mo ago
Yeh they seem to be in the same "major" cluster, although Serbian/Croatian, Romanian, Bulgarian, Turkish, Polish and Czech are all close. Turkish and Persian seem to be the nearest neighbors.
6.
by ilyausorov 11mo ago
Plotly is great! Much love.
7.
by ilyausorov 11mo ago
Yeh, we would've loved to see that too. It's on our roadmap for sure. Same for some of the other languages with a large amount of unique accents like e.g. French, Chinese, Arabic, etc...
8.
by ilyausorov 11mo ago
Nothing too secret in there! We anonymized everything and anyway it's just a basic Plotly plot. Feel free to check it out.
9.
by ilyausorov 11mo ago
Good question! It's likely because there are lots of different accents of Spanish that are distinct from each other. Our labels only capture the native language of the speaker right now, so they're all grouped together but it&#x27
10.
How AI hears accents: An audible visualization of accent clusters(accent-explorer.boldvoice.com)
260 points by ilyausorov 11mo ago|129 comments
11.
by ilyausorov 1y ago
Amazing! If you can make it go viral again too, I will love you!
12.
by ilyausorov 1y ago
We did built two free tools, which are geared towards non-native English speakers. You can find them at https://accentoracle.com and https://accentfilter.com . They're less effective for English native speakers,
13.
by ilyausorov 1y ago
Correct, not LLM
14.
by ilyausorov 1y ago
Indeed yeah that’s one of the key weaknesses of the approach that we’re using. It overrides the speakers cadence and accent while keeping their voice profile / timbre in place. Different techniques may not do this but also may not copy
15.
by ilyausorov 1y ago
Was that right? Or what is the correct native language it should have predicted? Note the %s in the accent breakdown section are prediction probabilities
16.
by ilyausorov 1y ago
What if it was already available? Try it out at https://accentfilter.com !
17.
by ilyausorov 1y ago
No, the dataset isn't published beyond what you see on the 2D visualization. Sorry.
18.
by ilyausorov 1y ago
Thanks, we're doing our best!
19.
by ilyausorov 1y ago
We actually did something like this for non-native English speakers a few months back. Check out https://accentoracle.com (most mind-blowing if you're a non native English speaker)
20.
by ilyausorov 1y ago
Indeed, although the inference output of the model is based on the ratings input that we trained it on. And that rating input was done by American English native speakers, so this iteration of the model is centered towards those accents mor
21.
by ilyausorov 1y ago
That's a fascinating idea! Definitely something to try out for our team. We actively and continuously do all sorts of experiments with our machine learning models to be able to extract the most useful insights. We will definitely share
22.
by ilyausorov 1y ago
Sure, that's fair. We apply labels that have a connotation of strength based on the distance, but the underlying calculation is indeed based on distance.
23.
by ilyausorov 1y ago
This and more exciting features are coming to the BoldVoice app soon!
24.
by ilyausorov 1y ago
For sure, and I don't think we ever use the term default or neutral. The "the American English accent of our expert accent coach Eliza" is just that -- it's one accent. As a learning platform that provides instruction to
25.
by ilyausorov 1y ago
Happy to see a happy BoldVoice user. Please don't hesitate to reach out to our team with feedback or thoughts on how we can continue to improve your learning journey. Helping you succeed is our #1 priority!
26.
by ilyausorov 1y ago
For sure we did! The training data we used for this was purposely highly varied to account for these various factors so they don't cause too much bias in the model. But there's also an error rate regardless of how good you make it
27.
by ilyausorov 1y ago
Fair point! When Victor tried to speed up to speak as fast as Coach Eliza, while it sounded somewhat less accented, a few parts of the phrase did get less intelligible. 10 minutes of practice is only a start after all. Interesting to note t
28.
Accents in latent spaces: How AI hears accent strength in English(accent-strength.boldvoice.com)
245 points by ilyausorov 1y ago|128 comments
29.
by ilyausorov 5y ago
Congrats on the launch guys! BoldVoice has been using GrowthBook for about ~6 weeks or so now, and it was super lucky that we found these guys right as we were considering way more expensive options like Optimizely (...6 figures). The tooli
30.
BoldVoice (YC S21) – Accent Coaching App – Is Hiring Full Stack Engineer(ycombinator.com)
1 points by ilyausorov 5y ago

More ›