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.

Generative Models

353 pointsby nicolapcweek94almost 9 years ago

13 comments

hasenjalmost 9 years ago
This is so cool and I can&#x27;t help but feel like I&#x27;m missing something important that&#x27;s taking place and has huge potential.<p>As a busy programmer who gets exhausted at night from the mental effort required at my day job, I have a feeling like I will never be able to catch up at this rate.<p>Are there any introductory materials to this field? Something I can read slowly during the weekends, that gives an overview of the fundamental concepts (primarily) and basic techniques (secondarily) without overwhelming the reader in the more advanced&#x2F;complicated techniques (at least during the beginning).<p>I&#x27;d really appreciate any recommendations.
评论 #11918790 未加载
评论 #11919902 未加载
评论 #11918623 未加载
评论 #11918717 未加载
评论 #11921665 未加载
评论 #11918629 未加载
评论 #11920807 未加载
评论 #11923665 未加载
andreykalmost 9 years ago
Brief summary: a nice intro about what generative models are and the current popular approaches&#x2F;papers, followed by descriptions of recent work by OpenAI in the space. Quick links to papers mentioned:<p>Improving GANs <a href="https:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;1606.03498" rel="nofollow">https:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;1606.03498</a><p>Improving VAEs <a href="http:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;1606.04934" rel="nofollow">http:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;1606.04934</a><p>InfoGAN <a href="https:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;1606.03657" rel="nofollow">https:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;1606.03657</a><p>Curiosity-driven Exploration in Deep Reinforcement Learning via Bayesian Neural Networks <a href="http:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;1605.09674" rel="nofollow">http:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;1605.09674</a><p>Generative Adversarial Imitation Learning <a href="http:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;1606.03476" rel="nofollow">http:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;1606.03476</a><p>I think the last one seems very exciting, I expect Imitation Learning would be a great approach for many robotics tasks.
brandonbalmost 9 years ago
Very cool. As you&#x27;re thinking about unsupervised or semi-supervised deep learning, consider medical data sets as a potential domain.<p>ImageNet has 1,034,908 labeled images. In a hospital setting, you&#x27;d be lucky to get 1000 participants.<p>That means those datasets really show off the power of unsupervised, semi-supervised, or one-shot learning algorithms. And if you set up the problem well, each increment of ROC translates into a life saved.<p>Happy to point you in the right direction when the time comes—my email is in my HN profile.
评论 #11918135 未加载
johnwatson11218almost 9 years ago
Have these techniques been used to generate realistic looking test data for testing software? I have had ideas along these lines but people think I&#x27;m talking about fuzz testing when I try and describe it.<p>I&#x27;m imagining something where you take a corporate db and reduce it down to a model. Then that can be shared with third parties and used to generate unlimited amounts of test data that looks like real data w&#x2F;o revealing any actual user info.
评论 #11921652 未加载
viachalmost 9 years ago
Looks like fake accounts on Facebook will have real unique userpics soon
评论 #11918081 未加载
评论 #11918072 未加载
ElHackeralmost 9 years ago
I really like that they used TensorFlow and published their code in GitHub. It will help a lot of people like me, that are new in the field and want to learn more about generative models. Amazing work by the OpenAI team!
评论 #11918366 未加载
bradscarletonalmost 9 years ago
It looks like they are using both TensorFlow and Theano. Is there a reason to use both?
评论 #11918091 未加载
j2kunalmost 9 years ago
The actual outputs look grotesque. Disembodied dog torsos with seven eyeballs and such. It&#x27;s cool, but to me this is clearly showing the local nature of convolutional nets; it&#x27;s a limitation that one has to overcome if one is to truly generate lifelike images from scratch.
评论 #11920825 未加载
评论 #11921690 未加载
dkarapetyanalmost 9 years ago
The generated images look like the stuff nightmares are made out of. Which is to say they&#x27;re extremely aesthetically unpleasant. So what exactly have these networks learned?
评论 #11920173 未加载
Rexxaralmost 9 years ago
Can we see somewhere the generated images with higher resolution ?
评论 #11919587 未加载
zumpalmost 9 years ago
Why do I constantly feel like I&#x27;m missing out with all this stuff?
pestaaalmost 9 years ago
What a beautifully presented research.
gradstudentalmost 9 years ago
Interesting topic, tedious article. Paraphrasing:<p>Q: What&#x27;s a generative model?<p>A: Well, we have these neural nets and...<p>Ugh. I understand the excitement for one&#x27;s own research but if the point is to make these results accessible to a wider audience then it&#x27;s important not to get lost in the details, at least not right away. IMO, there&#x27;s very little here in the way of high-level intuition. If I did not already have a PhD, and some exposure to ML (not my area), I would probably find this article entirely indecipherable. Again, paraphrasing:<p>Q: OK, so I understand you want to create pictures that resemble real photos. And you really like this DCGAN method, right?<p>A: Yes! See, it takes 100 random numbers and...<p>Come on guys. You can do better.
评论 #11921373 未加载
评论 #11919121 未加载
评论 #11920877 未加载