TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Zero-shot learning: Using text to more accurately identify images

117 点作者 pesenti超过 6 年前

2 条评论

askmike超过 6 年前
&gt; WHY IT MATTERS:<p>&gt; To become more useful, computer vision systems will need to recognize objects they have not specifically been trained on. For example, it is estimated that there are more than 10,000 living bird species, yet most computer vision data sets of birds have only a couple hundred categories.<p>I think it matters more for Facebook that they can apply these techniques to learn more about the 4 billion people that are not on Facebook, I don&#x27;t think they care that much about birds ;)
评论 #18405178 未加载
nl超过 6 年前
I&#x27;m sorry, but I work in this field and I find this so amazing that I don&#x27;t have any sensible comments to make.<p>I still find the fast.ai devise paper demo[1] amazing, and that is just working on labelled but unseen classes via the ImageNet hierarchy.<p>In this case they do really fine grained classification on unseen classes, with no hierarchy and instead exploit plain (noisy!) text.<p>This visual pivot thing is something I&#x27;m going to spend the next year coming back to I think.<p>[1] <a href="https:&#x2F;&#x2F;github.com&#x2F;fastai&#x2F;fastai&#x2F;blob&#x2F;master&#x2F;courses&#x2F;dl2&#x2F;devise.ipynb" rel="nofollow">https:&#x2F;&#x2F;github.com&#x2F;fastai&#x2F;fastai&#x2F;blob&#x2F;master&#x2F;courses&#x2F;dl2&#x2F;dev...</a>