最新评论 (0)


Latent Space Human Face Synthesis | Two Minute Papers #191

亲爱的学者朋友 这是 Károly Zsolnai-Fehér带来的两分钟论文
Dear Fellow Scholars, this is Two Minute Papers with Károly Zsolnai-Fehér.
In many previous episodes, we talked about generative adversarial networks, a recent
一个新方向 生成式对抗网络
new line in machine learning research with some
absolutely fantastic results in a variety of areas.
它们可以合成新的动物影相 从相片中创建三维模型
They can synthesize new images of animals, create 3d models from photos, or dream up
new products based on our edits of an image.
A generative adversarial network means that we have
two neural networks battling each other in an arms race.
这个生成网络旨在生成出更加真实的相片 然后交由
The generator network tries to create more and more realistic images, and these are passed
to the discriminator network which tries to learn the difference between real photographs
and fake, forged images.
在这个过程中 两个神经网络一起学习提高 直到变成自己
During this process, the two neural networks learn and improve together until they become
experts at their own craft.
如你所见 结果是非常振奋人心的
And as you can see, the results are fantastic.
However, training these networks against each other is anything but roses and sunshine.
我们不知道算法是否收敛 又或是达成纳什均衡
We don’t know if the process converges or if we reach Nash equilibrium.
Nash equilibrium is a state where both actors believe they have found an optimal strategy
可能的决定的情况下选取了最优策略的状态 且他们都不会
while taking into account the other actor’s possible decisions, and neither of them have
interest in changing their strategy.
这是博弈论中的经典场景 两个被定罪的罪犯正在考虑
This is a classical scenario in game theory where two convicted criminals are pondering
whether they should snitch on each other without knowing how the other decided to act.
如果你希望进一步了解纳什均衡 我在简介中已经放了可汗学院的
If you wish to hear more about the Nash-equilibrium, I’ve a put a link to Khan Academy’s video
视频链接 一定要去看看 你会喜欢的
in the description, make sure to check it out, you’ll love it!
在人工智能和博弈论中发现相似之处令我非常兴奋的 而更酷的是
I find it highly exciting that there are parallels in AI and game theory, however, the even cooler
thing is that here, we try to build a system where we don’t have to deal with such a situation.
这被称为生成潜在最优解 缩写为GLO 它用了一个小技巧
This is called Generative Latent Optimization, GLO in short and it is about introducing tricks
to do this by only using a generator network.
如果你曾经学习过字体设计 你就知道此领域是那么的复杂
If you have ever read up on font design, you know that it is a highly complex field.
但是 如果想要创建一种新的字体 我们的关注点一般
However, if we’d like to create a new font type, what we’re typically interested in is
仅集中在不多的几个特性上 比如他们的弯曲程度 是否有衬线
only a few features, like how curvy they are, or whether we’re dealing with a serif kind
of font, and simple descriptions like that.
此理论同样适用于人脸 动物 以及其它很多你能想到的主题
The same principle can be applied to human faces, animals, and most topics you can imagine.
This means that there are many complex concepts that contain a ton of information, most of
which can be captured by a simple description with only a few features.
This is done by projecting this high-dimensional data onto a low-dimensional latent space.
这个潜在空间帮助消除了对抗性优化 从而令整个系统
This latent space helps eliminating adversarial optimization, which makes this system much
更易训练 而它的主要卖点在于仍保留了生成式对抗网络
easier to train, and the main selling point is that it still retains the attractive properties
of generative adversarial networks.
This means that it can synthesize new samples from the learned dataset.
如果它学习了鸟的概念 它就能合成新的鸟类物种了
If it had learned the concept of birds, it will be able to synthesize new bird species.
It can perform continuous interpolation between data points.
这意味着 我们可以生成出两个选好的家具类型或灯具之间
This means that for instance, we can produce intermediate states between two chosen furniture
types or light fixtures.
It is also able to perform simple arithmetic operations between any number of data points.
比如 如果A群是带太阳镜的男性 B群是没有太阳眼镜的男性 C是
For instance, if A is males with sunglasses, B are males without sunglasses, and C are
女性 那么运算A-B+C就会生成戴太阳镜的女性
females, then A-B+C is going to generate females in sunglasses.
它也可以做到更清晰 更加更加清晰 记得一定要看看
It can also do super resolution and much, much more. Make sure to have a look at the
paper in the video description.
现在在结束之前 我们要指出一个大家避而不谈的问题:这些图像太小了
Now, before we go, we shall address the elephant in the room: these images are tiny.
经验丰富的学者们都知道 若要生成式对抗网络
Our seasoned Fellow Scholars know that for generative adversarial networks, there are
合成出包含更多细节的高分辨率图像 还有很多工作要做
plenty of works on how to synthesize high resolution images with more details.
这意味着 它指出了一个令人兴奋新方向
This means that this is a piece of work that opens up exciting new horizons, but it is
not to be measured against the tenth followup
work on top of a more established line of research.
两分钟论文会在此为您跟进 而这会正如
Two Minute Papers will be here for you to keep you updated on the progress, which is,
我们所知道的 机器学习研究的速度惊人的快
as we know, staggeringly quick in machine learning research.
别忘了订阅 点击提醒按钮以随时获取最新推送
Don’t forget to subscribe and click the bell icon to never miss an episode.
感谢您的观看和大力支持 我们下期再见!
Thanks for watching and for your generous support, and I’ll see you next time!