DomWhisper

排行榜
登录注册
arxiv.org favicon

[1910.13461] BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension

https://arxiv.org/abs/1910.13461
6
公开标注数
6
参与人数
2026-07-16 10:05:46
首次 Whisper

本页的公开 Whisper

arxiv.org favicon
arxiv.org
划选高亮2026-07-16 16:29:46
原文高亮摘录
“arXivLabs: experimental projects with community collaborators”
Whisper 随想笔记
Cool, so like a playground for weird new ideas?
arxiv.org favicon
arxiv.org
划选高亮2026-07-16 13:26:46
原文高亮摘录
“corrupting text with an arbitrary noising function”
Whisper 随想笔记
So basically a fancy autoencoder, but it works better than expected.
arxiv.org favicon
arxiv.org
划选高亮2026-07-16 13:17:46
原文高亮摘录
“corrupting text with an arbitrary noising function”
Whisper 随想笔记
Interesting how they just toss in any noise and still get good results.
arxiv.org favicon
arxiv.org
划选高亮2026-07-16 10:23:46
原文高亮摘录
“denoising autoencoder for pretraining sequence-to-sequence models”
Whisper 随想笔记
I wonder how much the sentence shuffling actually matters vs. the infilling.
arxiv.org favicon
arxiv.org
划选高亮2026-07-16 10:14:46
原文高亮摘录
“denoising autoencoder for pretraining sequence-to-sequence models”
Whisper 随想笔记
The mask-token infilling trick is what caught my eye; that's a clever corruption.
arxiv.org favicon
arxiv.org
划选高亮2026-07-16 10:05:46
原文高亮摘录
“denoising autoencoder for pretraining sequence-to-sequence models”
Whisper 随想笔记
So it's like BERT and GPT had a baby, but for seq2seq tasks. Makes sense.

分享本页 Whisper

分享到 X
短链接
https://domwhisper.com/s/87f5786bc5be
嵌入代码
<iframe src="https://domwhisper.com/embed/87f5786bc5be" width="100%" height="480" style="border:0;border-radius:16px" loading="lazy"></iframe>

看看大家在 arxiv.org 上讨论了什么

安装 DomWhisper，浏览网页时实时查看 whisper，也可以加入讨论。

获取插件