DomWhisper

Rankings
Log inSign Up
arxiv.org favicon

[1904.12848] Unsupervised Data Augmentation for Consistency Training

https://arxiv.org/abs/1904.12848
6
Public whispers
6
Contributors
2026-07-28 09:43:22
First whispered

Public whispers on this page

arxiv.org favicon
arxiv.org
Text Highlight2026-07-28 16:07:22
Original Highlight Excerpt
"arXivLabs: experimental projects with community collaborators"
Whisper Note
Is this just for the staff or can anyone pitch in?
arxiv.org favicon
arxiv.org
Text Highlight2026-07-28 13:04:22
Original Highlight Excerpt
"use of consistency training on a large amount of unlabe"
Whisper Note
So basically, they found that how you add noise matters more than the amount of unlabeled data?
arxiv.org favicon
arxiv.org
Text Highlight2026-07-28 12:55:22
Original Highlight Excerpt
"use of consistency training on a large amount of unlabe"
Whisper Note
Consistency training with better noise is the real winner here, not just the semi-supervised part.
arxiv.org favicon
arxiv.org
Text Highlight2026-07-28 10:01:22
Original Highlight Excerpt
"Semi-supervised learning lately has shown much promise in improving deep learning models when labeled data is scarce"
Whisper Note
We tried RandAugment in our pipeline and yeah, it made a bigger dent than I expected.
arxiv.org favicon
arxiv.org
Text Highlight2026-07-28 09:52:22
Original Highlight Excerpt
"Semi-supervised learning lately has shown much promise in improving deep learning models when labeled data is scarce"
Whisper Note
Still skeptical about those IMDb numbers with just 20 labels, but the idea is solid.
arxiv.org favicon
arxiv.org
Text Highlight2026-07-28 09:43:22
Original Highlight Excerpt
"Semi-supervised learning lately has shown much promise in improving deep learning models when labeled data is scarce"
Whisper Note
Finally someone pointing out that the noise itself matters, not just the consistency trick.

Share this page's whispers

Share to X
Short link
https://domwhisper.com/s/88fe27608f89
Embed snippet
<iframe src="https://domwhisper.com/embed/88fe27608f89" width="100%" height="480" style="border:0;border-radius:16px" loading="lazy"></iframe>

See what people are discussing on arxiv.org

Install DomWhisper to view live whispers as you browse, and join the discussion.

Get the extension