Wow, your work on jellyfish neural nets sounds so interesting. It is.

哇，你对水母神经网络的研究听起来好有趣。的确。

Are you really as suspicious of artificial neural nets as you pretend?

你真的这么不信任人工神经网络吗？

Now an AI can try and mimic that by abstracting generalizable behaviors using neural nets.

现在，人工智能可以尝试通过使用神经网络抽象出可概括的行为来模仿这一点。

But hydra have neural nets of diffuse brain cells, rather than a dense organ-like we have.

但水螅有弥漫的脑细胞构成的神经网络，而不像我们有一个密集的器官。

So I collected these drawings for the neural net. And we realized something that wasn't previously possible.

所以，我收集画作以训练神经网络，而且我们意识到前所未有的事情发生了。

By 2012, words like neural nets and machine learning were popping up on the front page of the New York Times.

2012 年，“神经网络”和“机器学习”等词汇涌现于纽约时报的头版之中。

The neural net does what the neural net does.

神经网络做神经网络所做的事情。

Hinton decides he's got an idea of how these neural nets might work, and he's going to pursue it no matter what.

辛顿确定他已经知道将如何使神经网络变得可行，便一以贯之 、追求不懈。

So I collected these drawings for the neural net.

所以我为神经网络收集了这些图纸。

They take their incoming data like we do from our senses, and they actually learn, so the neural net can learn to make decisions over time.

神经元像我们的感官，可以获取传入数据并进行学习，所以神经网络久而久之就能学会做决定。

And then in the mid-80s, we discovered how to make more complicated neural nets so they could solve those problems that the simple ones couldn't solve.

80 年代中期，我们发现了如何创建更复杂的神经网络，来解决简单的神经网络无法解决的问题。

We talked about neural nets last episode, if you need a primer.

我们上一集谈到了神经网络，如果你需要入门的话。

But neural nets require thousands of training examples to reliably form associations.

但是神经网络需要数以千计的训练样本才能可靠地形成关联。

I trained a neural net on these drawings in order to generate recurring patterns in the work that were then fed through custom software back into the machine.

我训练了一个神经网络来优化机器的循环模式，视觉样本由经专门的软件处理导入机器。

They're designed to work with a biological neural net, our brains, and with vision, our eyes.

它们旨在与生物神经网络（我们的大脑）和视觉（我们的眼睛）一起工作。

Now deep neural nets are driving cars, translating human speech, diagnosing medical conditions and much more.

现在，深度神经网络正在驾驶汽车、翻译人类语言、诊断医疗状况等等。

Another group's " neural physics engine" beat less structured neural nets at predicting ball collisions in containers.

另一组的“神经物理引擎”在预测容器中的球碰撞方面击败了结构化程度较低的神经网络。

This allowed it to solved CAPTCHAs better than other neural nets that were trained with 50,000 times more data.

这使其能够比其他接受 50,000 倍以上数据训练的神经网络更好地解决验证码问题。

So that's why our focus is so heavily on having a vision neural net

如果您无法处理可见性，则无法实现。

And I don't think we need a neural net running pattern-recognition software to see where this is all headed.

而且我认为我们不需要运行模式识别软件的神经网络来了解这一切的发展方向。