TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Show HN: OmniNet:- A unified architecture for multi-modal multi-task learning

54 pointsby amandavincialmost 6 years ago

7 comments

whoisnnamdialmost 6 years ago
Have not read the paper in much depth yet but this looks like great work, super interesting. Thanks for sharing.<p>Question: in the example of prediction on untrained tasks, what exactly hasn&#x27;t been trained? The paper talks about video being one of the trained tasks. Did you simply retrain model without video examples and then test performance?
评论 #20482361 未加载
absingh31almost 6 years ago
This one looks interesting, I think this paper would start multiple new researches towards AGI.
modi15almost 6 years ago
Can you explain in a bit layman terms what exactly has been done here ? What I understood is that there is a single NN trained for multiple tasks - but what is the benefit ?
pequalsnpalmost 6 years ago
I found my weekend paper. This sounds cool!
max_almost 6 years ago
Not an AI expert, but I wanted to know;<p>Does this bring us closer to AGI?
评论 #20480485 未加载
keerthankopalmost 6 years ago
Good work guys, well done.
mtmailalmost 6 years ago
Preferred original title when submitting links: &quot;Show HN: OmniNet - A unified architecture for multi-modal multi-task learning&quot;<p>HN is a bit strict. I&#x27;d say &quot;X is all you need&quot; gets less attention from users than a very technical headline. The most popular submission recently had MITM in the title (<a href="https:&#x2F;&#x2F;news.ycombinator.com&#x2F;best" rel="nofollow">https:&#x2F;&#x2F;news.ycombinator.com&#x2F;best</a>)
评论 #20479744 未加载