其他
【TED演讲】昆虫大脑是伟大人工智能的秘密吗？
TED英语演讲课
给心灵放个假吧
演讲题目：Are insect brains the secret to great AI?
演讲简介：
昆虫是大脑启发计算的关键吗？神经科学家弗朗西斯·S·钱斯是这样认为的。在这场有趣的讲座中，她分享了昆虫不可思议的能力的例子——比如蜻蜓极精确狩猎技能和非洲蜣螂的超强力量——并展示了解开它们微小大脑中神秘的神经元网络如何导致计算机、人工智能等领域的突破。
中英文字幕
在计算机上创建智能。
This has been the Holy Grail for artificial intelligence for quite some time.
很长一段时间以来，这一直是人工智能的圣杯。
But how do we get there?
但我们如何到达那里？
So we view ourselves as highly intelligent beings.
我们认为自己是高度智慧的人。
So it's logical to study our own brains, the substrate of our cognition, for creating artificial intelligence.
因此，研究我们自己的大脑,我们认知的基础,来创造人工智能是合乎逻辑的。
Imagine if we could replicate how our own brains work on a computer.
想象一下，如果我们可以在计算机上复制我们自己的大脑是如何工作的。
But now consider the journey that would be required.
但现在考虑一下所需的过程。
The human brain contains 86 billion neurons.
人类大脑包含860亿个神经元。
Each is constantly communicating with thousands of others.
每个人都在不断地与成千上万的人交流，
And each has individual characteristics of its own.
每个人都有自己的特点。
Capturing the human brain on a computer may simply be too big and too complex a problem,
在计算机上捕获人脑智慧可能的确是一个太大、太复杂的问题，
to tackle with the technology and the knowledge that we have today.
无法用我们今天的技术和知识来解决。
I believe that we can capture a brain on a computer.
我相信我们可以在计算机上捕获智慧，
But we have to start smaller.
但我们必须从更小的地方开始。
Much smaller.
小得多。
These insects have three of the most fascinating brains in the world to me.
对我来说，这些昆虫有三个世界上最迷人的大脑。
While they do not possess human-level intelligence, each is remarkable at a particular task.
虽然它们不具备人类水平的智力，但每一个都在特定活动中表现出色。
Think of them as highly trained specialists.
将他们视为训练有素的专家。
African dung beetles are really good at rolling large balls in straight lines.
非洲屎壳郎真的很擅长在直线上滚动大球。
Now, if you've ever made a snowman, you know that rolling a large ball is not easy.
如果你曾经堆过雪人，你就知道滚一个大球并不容易。
Now picture trying to make that snowman when the ball of snow is as big as you are.
现在想象一下堆雪人当雪球和你一样大时，
And you're standing on your head.
你倒立着。
Sahara desert ants are navigation specialists.
撒哈拉沙漠蚂蚁是导航专家。
They might have to wander a considerable distance to forage for food.
他们可能要走很远的路才能觅食。
But once they do find sustenance, they know how to calculate the straightest path home.
但一旦他们找到了食物，他们就知道如何计算回家的最直路径。
And the dragonfly is a hunting specialist.
而蜻蜓是狩猎专家。
In the wild, dragonflies capture approximately 95 percent of the prey they choose to go after.
在野外，蜻蜓捕获了大约95%的它们选择的猎物。
These insects are so good at their specialties that neuroscientists such as myself study them as model systems,
这些昆虫非常擅长它们的专业，以至于像我这样的神经科学家将它们作为模型系统来研究，
to understand how animal nervous systems solve particular problems.
以了解动物神经系统是如何解决特定的问题。
And in my own research, I study brains to bring these solutions, the best that biology has to offer, to computers.
在我的研究中，我研究大脑，以将这些生物所能提供的最好的解决方案引入计算机。
So consider the dragonfly brain.
想一下蜻蜓的大脑。
It has only on the order of one million neurons.
它只有大约100万个神经元。
Now, it's still not easy to unravel a circuit of even one million neurons.
现在，要解开一个哪怕有一百万个神经元的回路仍然不容易。
But given the choice between trying to tease apart the one-million-neuron brain versus the 86-billion-neuron brain,
但是如果要在尝试梳理100万个神经元大脑和860亿个神经元大脑之间做出选择，
which would you choose to try first?
你会选择先尝试哪一个？
When studying these smaller insect brains, the immediate goal is not human intelligence.
当研究这些较小的昆虫大脑时，当前的目标不是人类的智力。
We study these brains for what the insects do well.
我们研究这些大脑是为了了解昆虫做得好的地方。
And in the case of the dragonfly, that's interception.
就蜻蜓而言，那就是拦截。
So when dragonflies are hunting, they do more than just fly straight at the prey.
因此，当蜻蜓捕食时，它们所做的不仅仅是直接飞向猎物。
They fly in such a way that they will intercept it.
它们以这样的方式飞行，以拦截它。
They aim for where the prey is going to be.
它们瞄准猎物将要到达的地方。
Much like a soccer player, running to intercept a pass.
就像足球运动员，跑去拦截传球。
To do this correctly, dragonflies need to perform what is known as a coordinate transformation, going from the eye's frame of reference,
为了正确地做到这一点，蜻蜓需要进行所谓的坐标变换，从眼睛的参照系或蜻蜓看到的东西，
or what the dragonfly sees, to the body's frame of reference, or how the dragonfly needs to turn its body to intercept.
到身体的参照系，或者蜻蜓需要如何转动身体进行拦截。
Coordinate transformations are a basic calculation that animals need to perform to interact with the world.
坐标变换是动物与世界互动所需要进行的基本计算。
We do them instinctively every time we reach for something.
我们每次伸手拿东西的时候都会本能地做这些计算。
When I reach for an object straight in front of me, my arm takes a very different trajectory than if I turn my head,
当我伸手去拿我面前的一个物体时，
look at that same object when it is off to one side and reach for it there.
我的手臂的运动轨迹和我转头看向一边的同一物体时完全不同。
In both cases, my eyes see the same image of that object.
在这两种情况下，我的眼睛看到的都是同一物体的图像，
But my brain is sending my arm on a very different trajectory based on the position of my neck.
但我的大脑根据我脖子的位置将我的手臂送上一个非常不同的轨迹。
And dragonflies are fast.
蜻蜓很快。
This means they calculate fast.
这意味着他们计算得很快。
The latency, or the time it takes for a dragonfly to respond once it sees the prey turn, is about 50 milliseconds.
延迟，即蜻蜓在看到猎物转向后做出反应所需的时间，大约是50毫秒。
This latency is remarkable.
这种延迟是很了不起的。
For one thing, it's only half the time of a human eye blink.
一方面，这只是人类眨眼时间的一半。
But for another thing, it suggests that dragonflies capture how to intercept in only relatively or surprisingly few computational steps.
但另一方面，它表明蜻蜓仅通过相对的或惊人的极少计算步骤即可体现出如何进行拦截。
So in the brain, a computational step is a single neuron or a layer of neurons working in parallel.
所以在大脑中，计算步骤是单个神经元或一层神经元并行工作。
It takes a single neuron about 10 milliseconds to add up all its inputs and respond.
单个神经元需要大约10毫秒才能将其所有输入相加并做出反应。
The 50-millisecond response time means that once the dragonfly sees its prey turn,
50毫秒的响应时间意味着，一旦蜻蜓看到它的猎物转向，
there's only time for maybe four of these computational steps or four layers of neurons, working in sequence, one after the other,
可能只有四个计算步骤或四层神经元依次工作的时间，一个接一个，
to calculate how the dragonfly needs to turn.
来计算蜻蜓需要如何转向。
In other words, if I want to study how the dragonfly does coordinate transformations, the neural circuit that I need to understand,
换句话说，如果我想研究蜻蜓如何进行坐标变换，我需要了解神经回路，
the neural circuit that I need to study, can have at most four layers of neurons.
我需要研究神经回路，最多可以有四层神经元。
Each layer may have many neurons.
每一层可能有许多神经元，
But this is a small neural circuit.
但这是一个小的神经回路。
Small enough that we can identify it and study it with the tools that are available today.
小到我们可以用今天的工具来识别它和研究它。
And this is what I'm trying to do.
这就是我要做的。
I have built a model of what I believe is the neural circuit that calculates how the dragonfly should turn.
我已经建立了一个我认为是计算蜻蜓应该如何转向的神经回路的模型。
And here is the cool result.
这是一个很酷的结果。
In the model, dragonflies do coordinate transformations in only one computational step, one layer of neurons.
在该模型中，蜻蜓只用一个计算步骤，即一个神经元层来做坐标转换。
This is something we can test and understand.
这是我们可以测试和理解的。
In a computer simulation, I can predict the activities of individual neurons while the dragonfly is hunting.
在计算机模拟中，我可以预测蜻蜓狩猎时单个神经元的活动。
For example, here I am predicting the action potentials, or the spikes, that are fired by one of these neurons when the dragonfly sees the prey move.
例如，我在这里预测当蜻蜓看到猎物移动时，其中一个神经元发射了动作电位或脉冲。
To test the model,
为了测试这个模型，
my collaborators and I are now comparing these predicted neural responses with responses of neurons recorded in living dragonfly brains.
我和我的合作者现在正在将这些预测的神经反应与活体蜻蜓大脑中记录的神经元反应进行比较。
These are ongoing experiments in which we put living dragonflies in virtual reality.
这些是正在进行的实验，我们将活体蜻蜓放在虚拟现实中。
Now, it's not practical to put VR goggles on a dragonfly.
现在，给蜻蜓戴上VR护目镜是不现实的。
So instead, we show movies of moving targets to the dragonfly, while an electrode records activity patterns of individual neurons in the brain.
因此，我们改为向蜻蜓播放移动目标的电影，同时电极记录大脑中单个神经元的活动模式。
Yeah, he likes the movies.
是的，他喜欢电影。
If the responses that we record in the brain match those predicted by the model,
如果我们在大脑中记录的反应与模型预测的反应相匹配，
we will have identified which neurons are responsible for coordinate transformations.
我们就会确定哪些神经元负责坐标转换。
The next step will be to understand the specifics of how these neurons work together to do the calculation.
下一步将是了解这些神经元如何协同工作进行计算的细节。
But this is how we begin to understand how brains do basic or primitive calculations.
但这就是我们开始了解大脑是如何进行基本或原始的计算。
Calculations that I regard as building blocks for more complex functions, not only for interception but also for cognition.
计算，我将其视为更复杂功能的构件，不仅用于拦截，还用于认知。
The way that these neurons compute may be different from anything that exists on a computer today.
这些神经元的计算方式可能不同于当今计算机上存在的任何东西。
And the goal of this work is to do more than just write code that replicates the activity patterns of neurons.
这项工作的目标不仅仅是编写复制神经元活动模式的代码。
We aim to build a computer chip that not only does the same things as biological brains but does them in the same way as biological brains.
我们的目标是制造一种计算机芯片，它不仅可以做与生物大脑相同的事情，而且可以用与生物大脑同样的方式来做这些事情。
This could lead to drones driven by computers the same size of the dragonfly's brain, that captures some targets and avoid others.
这可能会导致由计算机驱动的无人机，其大小与蜻蜓的大脑相同，捕获一些目标并避开其他目标。
Personally, I'm hoping for a small army of these to defend my backyard from mosquitoes in the summer.
就我个人而言，我希望有一小群这样的无人机在夏天保护我的后院不受蚊子骚扰。
The GPS on your phone could be replaced by a new navigation device based on dung beetles or ants,
你手机上的GPS可能会被一种基于蜣螂或蚂蚁的新型导航设备所取代，
that could guide you to the straight or the easy path home.
它可以引导你走直路或容易回家的路。
And what would the power requirements of these devices be like?
那么这些设备的功率要求是怎样的呢？
As small as it is...
尽管它很小……
Or, sorry, as large as it is, the human brain is estimated to have the same power requirements as a 20-watt light bulb.
或者说，对不起，尽管它很大，据估计，人脑的功率需求与20瓦的灯泡相同。
Imagine if all brain-inspired computers had the same extremely low-power requirements.
想象一下，如果所有受大脑启发的计算机都具有相同的极低功耗要求。
Your smartphone or your smartwatch probably needs charging every day.
你的智能手机或智能手表可能每天都需要充电。
Your new brain-inspired device might only need charging every few months, or maybe even every few years.
你的新大脑启发设备可能只需要每隔几个月，甚至几年充电一次。
The famous physicist, Richard Feynman, once said: What I cannot create, I do not understand.
著名物理学家理查德·费曼曾说：“我不能创造的东西，我就不了解。”
What I see in insect nervous systems is an opportunity to understand brains through the creation of computers that work as brains do.
我在昆虫神经系统中看到的是一个通过创造与大脑一样工作的计算机来了解大脑的机会。
And creation of these computers will not just be for knowledge.
而这些计算机的创造将不仅仅是为了认知。
There's potential for real impact on your devices, your vehicles, maybe even artificial intelligences.
有可能对你的设备、车辆甚至是人工智能产生真正的影响。
So next time you see an insect, consider that these tiny brains can lead to remarkable computers.
所以，下次你看到一只昆虫时，想想看，这些微小的大脑可以发展出卓越的计算机。
And think of the potential that they offer us for the future.
想想它们为我们的未来提供的潜力。
Thank you.
谢谢。
视频、演讲稿均来源于TED官网