AI Research Lessons: Computation Key

The biggest lesson that can be read from 70 years of AI research is that general methods that leverage computation are ultimately the most effective, and by a large margin.

The biggest lesson that can be read from 70 years of AI research is that general methods that leverage computation are ultimately the most effective, and by a large margin.

可以从70年的人工智能研究中得出的最大教训是，利用计算的一般方法最终是最有效的，而且优势巨大。

The biggest lesson that can be read from 70 years of AI research is that general methods that leverage computation are ultimately the most effective, and by a large margin. The ultimate reason for this is Moore's law, or rather its generalization of continued exponentially falling cost per unit of computation. Most AI research has been conducted as if the computation available to the agent were constant (in which case leveraging human knowledge would be one of the only ways to improve performance) but, over a slightly longer time than a typical research project, massively more computation inevitably becomes available. Seeking an improvement that makes a difference in the shorter term, researchers seek to leverage their human knowledge of the domain, but the only thing that matters in the long run is the leveraging of computation. These two need not run counter to each other, but in practice they tend to. Time spent on one is time not spent on the other. There are psychological commitments to investment in one approach or the other. And the human-knowledge approach tends to complicate methods in ways that make them less suited to taking advantage of general methods leveraging computation. There were many examples of AI researchers' belated learning of this bitter lesson, and it is instructive to review some of the most prominent.

The biggest lesson that can be read from 70 years of AI research is that general methods that leverage computation are ultimately the most effective, and by a large margin.

可以从70年的人工智能研究中得出的最大教训是，利用计算的一般方法最终是最有效的，而且优势巨大。

The ultimate reason for this is Moore's law, or rather its generalization of continued exponentially falling cost per unit of computation.

这背后的根本原因是摩尔定律，或者更确切地说，是持续指数级降低的计算单位成本的普遍化。

Most AI research has been conducted as if the computation available to the agent were constant (in which case leveraging human knowledge would be one of the only ways to improve performance) but, over a slightly longer time than a typical research project, massively more computation inevitably becomes available.

大多数人工智能研究都是在假设可用于代理的计算资源是恒定的情况下进行的（在这种情况下，利用人类知识将是提高性能的唯一途径之一），但在比一般研究项目略长的时间内，大量更多的计算资源不可避免地会变得可用。

Seeking an improvement that makes a difference in the shorter term, researchers seek to leverage their human knowledge of the domain, but the only thing that matters in the long run is the leveraging of computation.

为了寻求短期内有所改进的方法，研究人员试图利用他们对领域的人类知识，但从长远来看，唯一重要的是利用计算。

These two need not run counter to each other, but in practice they tend to.

这两者不必相互对立，但实际上它们往往如此。

Time spent on one is time not spent on the other.

在一个方面花费的时间就是没有在另一个方面花费的时间。

There are psychological commitments to investment in one approach or the other.

在一种方法或另一种方法上的投资存在心理上的承诺。

And the human-knowledge approach tends to complicate methods in ways that make them less suited to taking advantage of general methods leveraging computation.

而人类知识方法往往以复杂化方法的方式进行，使它们不太适合利用利用计算的一般方法。

There were many examples of AI researchers' belated learning of this bitter lesson, and it is instructive to review some of the most prominent.

人工智能研究人员迟来的学习这个苦涩教训的例子很多，回顾其中一些最突出的例子是具有启发性的。

In computer chess, the methods that defeated the world champion, Kasparov, in 1997, were based on massive, deep search. At the time, this was looked upon with dismay by the majority of computer-chess researchers who had pursued methods that leveraged human understanding of the special structure of chess. When a simpler, search-based approach with special hardware and software proved vastly more effective, these human-knowledge-based chess researchers were not good losers. They said that ``brute force" search may have won this time, but it was not a general strategy, and anyway it was not how people played chess. These researchers wanted methods based on human input to win and were disappointed when they did not.

In computer chess, the methods that defeated the world champion, Kasparov, in 1997, were based on massive, deep search.

在计算机象棋中，1997年击败世界冠军卡斯帕罗夫的方法是基于大规模深度搜索。

At the time, this was looked upon with dismay by the majority of computer-chess researchers who had pursued methods that leveraged human understanding of the special structure of chess.

当时，这被大多数追求利用人类对象棋特殊结构理解的计算机象棋研究人员以惊愕的眼光看待。

When a simpler, search-based approach with special hardware and software proved vastly more effective, these human-knowledge-based chess researchers were not good losers.

当一个更简单、基于搜索的方法配合特殊的硬件和软件证明远远更有效时，这些基于人类知识的象棋研究者并没有好好接受失败。

They said that ``brute force" search may have won this time, but it was not a general strategy, and anyway it was not how people played chess.

他们说，“暴力”搜索可能这次赢了，但它不是一种通用策略，而且无论如何这不是人们下棋的方式。

These researchers wanted methods based on human input to win and were disappointed when they did not.

这些研究人员希望基于人类输入的方法能够获胜，并在它们没有获胜时感到失望。

A similar pattern of research progress was seen in computer Go, only delayed by a further 20 years. Enormous initial efforts went into avoiding search by taking advantage of human knowledge, or of the special features of the game, but all those efforts proved irrelevant, or worse, once search was applied effectively at scale. Also important was the use of learning by self play to learn a value function (as it was in many other games and even in chess, although learning did not play a big role in the 1997 program that first beat a world champion). Learning by self play, and learning in general, is like search in that it enables massive computation to be brought to bear. Search and learning are the two most important classes of techniques for utilizing massive amounts of computation in AI research. In computer Go, as in computer chess, researchers' initial effort was directed towards utilizing human understanding (so that less search was needed) and only much later was much greater success had by embracing search and learning.

A similar pattern of research progress was seen in computer Go, only delayed by a further 20 years.

计算机围棋的研究进展出现了类似的模式，只是延迟了20年。

Enormous initial efforts went into avoiding search by taking advantage of human knowledge, or of the special features of the game, but all those efforts proved irrelevant, or worse, once search was applied effectively at scale.

最初巨大的努力都用于避免搜索，利用人类知识或游戏的特殊特征，但一旦在大规模上有效应用搜索，所有这些努力都被证明是无关紧要的，甚至更糟。

Also important was the use of learning by self play to learn a value function (as it was in many other games and even in chess, although learning did not play a big role in the 1997 program that first beat a world champion).

同样重要的是通过自我对弈学习价值函数的使用（就像在许多其他游戏甚至是象棋中一样，尽管学习在1997年首次击败世界冠军的程序中并没有发挥重要作用）。

Learning by self play, and learning in general, is like search in that it enables massive computation to be brought to bear.

通过自我对弈学习，以及一般的学习，就像搜索一样，它使得可以利用大规模的计算。

Search and learning are the two most important classes of techniques for utilizing massive amounts of computation in AI research.

搜索和学习是利用大量计算进行人工智能研究的两个最重要的技术类别。

In computer Go, as in computer chess, researchers' initial effort was directed towards utilizing human understanding (so that less search was needed) and only much later was much greater success had by embracing search and learning.

在计算机围棋中，就像在计算机象棋中一样，研究人员最初的努力是为了利用人类的理解（从而减少搜索的需要），直到很久以后，通过拥抱搜索和学习才取得了更大的成功。

In speech recognition, there was an early competition, sponsored by DARPA, in the 1970s. Entrants included a host of special methods that took advantage of human knowledge---knowledge of words, of phonemes, of the human vocal tract, etc. On the other side were newer methods that were more statistical in nature and did much more computation, based on hidden Markov models (HMMs). Again, the statistical methods won out over the human-knowledge-based methods. This led to a major change in all of natural language processing, gradually over decades, where statistics and computation came to dominate the field. The recent rise of deep learning in speech recognition is the most recent step in this consistent direction. Deep learning methods rely even less on human knowledge, and use even more computation, together with learning on huge training sets, to produce dramatically better speech recognition systems. As in the games, researchers always tried to make systems that worked the way the researchers thought their own minds worked---they tried to put that knowledge in their systems---but it proved ultimately counterproductive, and a colossal waste of researcher's time, when, through Moore's law, massive computation became available and a means was found to put it to good use.

In speech recognition, there was an early competition, sponsored by DARPA, in the 1970s.

在语音识别领域，1970年代有一个早期的由DARPA赞助的比赛。

Entrants included a host of special methods that took advantage of human knowledge---knowledge of words, of phonemes, of the human vocal tract, etc.

参赛者包括了许多利用人类知识的特殊方法——单词知识、音素知识、人类发音道知识等。

On the other side were newer methods that were more statistical in nature and did much more computation, based on hidden Markov models (HMMs).

另一方面是性质更统计学的新方法，它们进行了更多的计算，基于隐藏马尔可夫模型（HMMs）。

Again, the statistical methods won out over the human-knowledge-based methods.

再次，统计方法胜过了基于人类知识的方法。

This led to a major change in all of natural language processing, gradually over decades, where statistics and computation came to dominate the field.

这导致了自然语言处理领域的一次重大变革，在几十年的时间里，统计和计算逐渐主导了这个领域。

The recent rise of deep learning in speech recognition is the most recent step in this consistent direction.

深度学习在语音识别中的最近兴起是这一持续方向的最新步骤。

Deep learning methods rely even less on human knowledge, and use even more computation, together with learning on huge training sets, to produce dramatically better speech recognition systems.

深度学习方法更少地依赖人类知识，使用更多的计算，并结合在大型训练集上的学习，以产生大幅度提升的语音识别系统。

As in the games, researchers always tried to make systems that worked the way the researchers thought their own minds worked---they tried to put that knowledge in their systems---but it proved ultimately counterproductive, and a colossal waste of researcher's time, when, through Moore's law, massive computation became available and a means was found to put it to good use.

就像在游戏中一样，研究人员总是试图制造出按照他们认为自己的思维方式工作的系统——他们试图将这种知识放入他们的系统——但最终证明这是适得其反的，而且是研究人员时间的巨大浪费，因为通过摩尔定律，大量的计算资源变得可用，并找到了将其有效利用的方法。

In computer vision, there has been a similar pattern. Early methods conceived of vision as searching for edges, or generalized cylinders, or in terms of SIFT features. But today all this is discarded. Modern deep-learning neural networks use only the notions of convolution and certain kinds of invariances, and perform much better.

This is a big lesson. As a field, we still have not thoroughly learned it, as we are continuing to make the same kind of mistakes. To see this, and to effectively resist it, we have to understand the appeal of these mistakes. We have to learn the bitter lesson that building in how we think we think does not work in the long run. The bitter lesson is based on the historical observations that 1) AI researchers have often tried to build knowledge into their agents, 2) this always helps in the short term, and is personally satisfying to the researcher, but 3) in the long run it plateaus and even inhibits further progress, and 4) breakthrough progress eventually arrives by an opposing approach based on scaling computation by search and learning. The eventual success is tinged with bitterness, and often incompletely digested, because it is success over a favored, human-centric approach.

In computer vision, there has been a similar pattern.

在计算机视觉领域，出现了类似的模式。

Early methods conceived of vision as searching for edges, or generalized cylinders, or in terms of SIFT features.

早期方法将视觉理解为寻找边缘、广义圆柱体或是基于SIFT特征。

But today all this is discarded.

但今天这一切都被抛弃了。

Modern deep-learning neural networks use only the notions of convolution and certain kinds of invariances, and perform much better.

现代深度学习神经网络只使用卷积的概念和某些类型的不变性，并且表现得更好。

This is a big lesson.

这是一个重要的教训。

As a field, we still have not thoroughly learned it, as we are continuing to make the same kind of mistakes.

作为一个领域，我们还没有彻底学会这个教训，因为我们还在继续犯同样的错误。

To see this, and to effectively resist it, we have to understand the appeal of these mistakes.

要看到这一点，并有效地抵制它，我们必须理解这些错误的吸引力。

We have to learn the bitter lesson that building in how we think we think does not work in the long run.

我们必须学会这个苦涩的教训：构建我们认为我们思考的方式从长远来看是行不通的。

The bitter lesson is based on the historical observations that 1) AI researchers have often tried to build knowledge into their agents, 2) this always helps in the short term, and is personally satisfying to the researcher, but 3) in the long run it plateaus and even inhibits further progress, and 4) breakthrough progress eventually arrives by an opposing approach based on scaling computation by search and learning.

这个苦涩的教训是基于历史观察：1) 人工智能研究人员经常试图将知识构建进他们的代理中，2) 这在短期内总是有帮助的，并且对研究人员个人来说是满足的，但3) 从长远来看，它会达到平台期甚至阻碍进一步的进展，4) 突破性进展最终是通过基于搜索和学习的计算扩展的对立方法到来的。

The eventual success is tinged with bitterness, and often incompletely digested, because it is success over a favored, human-centric approach.

最终的成功带有苦涩的味道，而且经常消化不良，因为它是在受到青睐的以人为中心的方法上的成功。

One thing that should be learned from the bitter lesson is the great power of general purpose methods, of methods that continue to scale with increased computation even as the available computation becomes very great. The two methods that seem to scale arbitrarily in this way are search and learning.

The second general point to be learned from the bitter lesson is that the actual contents of minds are tremendously, irredeemably complex; we should stop trying to find simple ways to think about the contents of minds, such as simple ways to think about space, objects, multiple agents, or symmetries. All these are part of the arbitrary, intrinsically-complex, outside world. They are not what should be built in, as their complexity is endless; instead we should build in only the meta-methods that can find and capture this arbitrary complexity. Essential to these methods is that they can find good approximations, but the search for them should be by our methods, not by us. We want AI agents that can discover like we can, not which contain what we have discovered. Building in our discoveries only makes it harder to see how the discovering process can be done.

One thing that should be learned from the bitter lesson is the great power of general purpose methods, of methods that continue to scale with increased computation even as the available computation becomes very great.

从苦涩的教训中应学到的一件事是通用方法的巨大力量，即随着可用计算量的增加而持续扩展的方法。

The two methods that seem to scale arbitrarily in this way are search and learning.

以这种方式似乎可以任意扩展的两种方法是搜索和学习。

The second general point to be learned from the bitter lesson is that the actual contents of minds are tremendously, irredeemably complex; we should stop trying to find simple ways to think about the contents of minds, such as simple ways to think about space, objects, multiple agents, or symmetries.

从苦涩的教训中学到的第二个一般性观点是，心灵的实际内容是极其复杂且无可挽回的；我们应该停止尝试寻找思考心灵内容的简单方式，比如思考空间、物体、多个代理或对称性的简单方式。

All these are part of the arbitrary, intrinsically-complex, outside world.

所有这些都是任意的、本质上复杂的、外部世界的一部分。

They are not what should be built in, as their complexity is endless; instead we should build in only the meta-methods that can find and capture this arbitrary complexity.

它们不是应该内置的东西，因为它们的复杂性是无穷的；相反，我们应该只构建能够找到并捕捉这种任意复杂性的元方法。

Essential to these methods is that they can find good approximations, but the search for them should be by our methods, not by us.

这些方法的关键是它们能够找到良好的近似值，但寻找它们应该通过我们的方法，而不是我们自己。

We want AI agents that can discover like we can, not which contain what we have discovered.

我们希望有能像我们一样发现的人工智能代理，而不是包含我们所发现的内容。

Building in our discoveries only makes it harder to see how the discovering process can be done.

内置我们的发现只会使看到发现过程如何完成变得更加困难。