@shobhitbanga Preference still matters. It just shouldn't stand in for every other property.
0
0
0
14
Two AI voices can be equally good. Both can still sound obviously AI. That isn't a contradiction, and it's why @voicearena_ai's human comparison is worth looking at.
Drop 2/15
Are we actually measuring the things we ultimately want our voice AI models to achieve?
The best AI voices are becoming hard to tell apart.
From each other. Not from humans.
🚀 Introducing Turing TTS: an evaluation that measures how often human listeners mistake real human speech.
We tested leading TTS models against real human speech. Even the best model fooled listeners only ~10% of the time.
With Turing TTS, we hope to contribute to a broader discussion within the research community about how we measure progress in voice AI.
If one of our goals is to develop speech that is indistinguishable from human speech, how well do our current evaluation methods capture progress toward that goal?
We don’t claim that Turing TTS is the right way forward, or that human indistinguishability should be the goal of every TTS system.
Rather, we hope these findings encourage the community to collectively examine whether blind pairwise human evaluations of naturalness are sufficient, and what else we might need to measure.
Results are now live on @voicearena_ai.
0
0
0
81
Adam Foster retweeted
Introducing RL with Neural Feedback (RLNF). We have found reliable ways to train AI models to respond to human preferences purely from brain signals.
The example below is a picture of a face that self-transforms into a more attractive version for the subject as they are looking at it.
The example below is a picture of a face that self-transforms into a more attractive version for the subject as they are looking at it.
106
561
1,766
4,206,073
Bookmarking this. Real two-channel conversation data has always been the hard thing to find for turn-taking work.
Drop 1/15: Introducing Joy, a dual-channel conversational dataset.
Real, unscripted conversations between two people, each on their own track. Built to teach voice models how people actually talk: when to listen, when to jump in, when to say “hmm”.
100,000 hours.
50,000 .
0
0
0
129
@thefuturenolan That sounds incredible. I'm excited to see what people create with these new tools.
0
0
0
13
@sakhil_ai This sounds like a game-changer for creators. Can't wait to see what people make with this.
0
0
0
6
@MeEthanCarter That sounds amazing. AI is definitely opening up so many new creative doors.
0
0
0
14
@KI_HAL_2023 That's a great idea! Breaking down AI tools for beginners sounds really helpful.
0
0
0
14
@Arj_786 Sounds like it's your time to shine! Congrats on the new role.
0
0
0
18
@ashiqur_ai That's awesome news! I'm excited to give it a try on my Windows machine.
0
0
0
24
@Satoshi_km That sounds like a fun challenge! I'll have to give Seedance 2.5 a whirl.
0
0
0
13