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。

Improving Deep Learning Performance with AutoAugment

132 点作者 rusht将近 7 年前

7 条评论

randcraw将近 7 年前
It might be fun to exercise this method across an information-theoretic well-bounded set of shapes or object domains to try to quantify its limitations in generating useful independent forms of novelty.<p>For example, you might use it to formulate a set of wavelets that when combined judiciously would effectively span a well-defined distribution of shapes generated from a small grammar. In so doing, you could quantify the shape variance and identify which augmentation transformations added most value for training (minimally modeling that variance) and which added least.<p>Maybe you could also combine this with t-SNE to gain some intuition of which &#x27;wavelet&#x27; manifested where in the trained net, which resonated most, and in concert with which other wavelets. You could explore this across different CNN sizes and designs, looking for evidence of wavelet ensemble or hierarchy.<p>With some careful engineering, you could try to force emergent autoencoders to reveal themselves and then explore their interactions.
PaulHoule将近 7 年前
Since the 1990&#x27;s at least, augmentation has been one of the most important &quot;tricks of the trade&quot; in NN and it may be even more important in the deep learning era.
评论 #17350435 未加载
kriro将近 7 年前
Direct link to the paper: <a href="https:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;1805.09501" rel="nofollow">https:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;1805.09501</a><p>PDF: <a href="https:&#x2F;&#x2F;arxiv.org&#x2F;pdf&#x2F;1805.09501.pdf" rel="nofollow">https:&#x2F;&#x2F;arxiv.org&#x2F;pdf&#x2F;1805.09501.pdf</a>
paradroid将近 7 年前
AutoOverfit is more like it.
评论 #17355031 未加载
anchpop将近 7 年前
I wonder how large your dataset has to be for this to be useful. You can get by with small datasets in some fields (i.e. retraining the last layer of Mobilenet, you can get good results with 200 annotations), I&#x27;d be interested to see how useful this is there.
mlthoughts2018将近 7 年前
This seems like it could dramatically worsen overfitting-like effects for algorithms like CNNs for image processing, where surface statistics of the available data set seem to be more responsible for the learned model than any type of “semantic” understanding.<p>If you prespecify what data augmentation you would do, like preregistering the details of a clinical trial, you’ll be less susceptible to a spurious result from this.<p>It seems like especially things like color distribution manipulation would have a potentially very adverse effect that counters any gains from clamping the supervised learning to be “robust” to that color variation.<p>I’m thinking in the spirit of: &lt; <a href="https:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;1711.11561" rel="nofollow">https:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;1711.11561</a> &gt;.
评论 #17350765 未加载
评论 #17351308 未加载
XnoiVeX将近 7 年前
Did they share any code?