【论文翻译】Deep learning

论文题目：Deep Learning
论文来源:Deep Learning_2015_Nature

Deep Learning
Yann LeCun∗ Yoshua Bengio∗ Geoffrey Hinton
深度学习
Yann LeCun∗ Yoshua Bengio∗ Geoffrey Hinton

Abstract

Deep learning allows computational models that are composed of multiple processing layers to learn representations of data with multiple levels of abstraction. These methods have dramatically improved the state-of-the-art in speech rec- ognition, visual object recognition, object detection and many other domains such as drug discovery and genomics. Deep learning discovers intricate structure in large data sets by using the backpropagation algorithm to indicate how a machine should change its internal parameters that are used to compute the representation in each layer from the representation in the previous layer. Deep convolutional nets have brought about breakthroughs in processing images, video, speech and audio, whereas recurrent nets have shone light on sequential data such as text and speech.

摘要

深度学习允许由多个处理层组成的计算模型学习具有多个抽象级别的数据表示。这些方法极大地改善了语音识别，视觉对象识别，对象检测以及许多其他领域的最新技术，例如药物发现和基因组学。深度学习通过使用反向传播算法来指示机器应如何更改其内部参数（从上一层的表示形式计算每一层的表示形式）中，从而发现大型数据集中的复杂结构。深层卷积网络在处理图像，视频，语音和音频方面有所突破，而递归网络则对处理诸如文本和语音之类的顺序数据中有应用前景。

正文

Machine-learning technology powers many aspects of modern society: from web searches to content filtering on social net- works to recommendations on e-commerce websites, andit is increasingly present in consumer products such as cameras and smartphones. Machine-learning systems are used to identify objects in images, transcribe speech into text, match news items, posts or products with users’ interests, and select relevant results of search. Increasingly, these applications make use of a class of techniques called deep learning.
机器学习技术为现代社会的各个方面提供了强大的便利：从网络搜索到社交网络上的内容过滤，再到电子商务网站上的推荐，机器学习技术越来越多地出现在诸如照相机和智能手机之类的消费产品中。机器学习系统用于识别图像中的对象，将语音转换为文本，使得新闻，帖子或具有用户兴趣的产品相匹配，以及选择相关的搜索结果。这些应用程序越来越多地使用了深度学习技术。
Conventional machine-learning techniques were limited in their ability to process natural data in their raw form. For decades, con- structing a pattern-recognition or machine-learning system required careful engineering and considerable domain expertise to design a fea- ture extractor that transformed the raw data (such as the pixel values of an image) into a suitable internal representation or feature vector from which the learning subsystem, often a classifier, could detect or classify patterns in the input.
传统的机器学习技术在处理原始格式的自然数据方面受到限制。几十年来，构建模式识别或机器学习系统需要认真的工程设计和相当多的领域专业知识，才能设计特征提取器，以将原始数据（例如图像的像素值）转换为合适的内部表示形式或特征向量，学习子系统（通常是分类器）可以根据特征向量检测或分类输入中的模式。
Representation learning is a set of methods that allows a machine to be fed with raw data and to automatically discover the representations needed for detection or classification. Deep-learning methods are representation-learning methods with multiple levels of representa- tion, obtained by composing simple but non-linear modules that each transform the representation at one level (starting with the raw input) into a representation at a higher, slightly more abstract level. With the composition of enough such transformations, very complex functions can be learned. For classification tasks, higher layers of representation amplify aspects of the input that are important for discrimination and suppress irrelevant variations. An image, for example, comes in the form of an array of pixel values, and the learned features in the first layer of representation typically represent the presence or absence of edges at particular orientations and locations in the image. The second layer typically detects motifs by spotting particular arrangements of edges, regardless of small variations in the edge positions. The third layer may assemble motifs into larger combinations that correspond to parts of familiar objects, and subsequent layers would detect objects as combinations of these parts. The key aspect of deep learning is that these layers of features are not designed by human engineers: they are learned from data using a general-purpose learning procedure.
表示学习是一组允许向机器提供原始数据并自动发现检测或分类所需的表示的方法。深度学习方法是具有多层表示形式的表示学习方法，它是通过组合简单但非线性的模块而获得的，每个模块都将一个级别的表示（从原始输入开始）转换为更高，稍高一点的表示。抽象级别。有了足够多的此类转换，就可以学习非常复杂的功能。对于分类任务，较高的表示层会放大输入中对区分非常重要的方面，并抑制不相关的变化。例如，图像以像素值阵列的形式出现，并且在表示的第一层中学习的特征通常表示图像中特定方向和位置上是否存在边缘。第二层通常通过发现边缘的特定布置来检测图案，而与边缘位置的微小变化无关。第三层可以将图案组装成与熟悉的对象的各个部分相对应的较大组合，并且随后的层将对象检测为这些部分的组合。深度学习的关键方面是这些层的功能不是由人类工程师设计的：它们是使用通用学习过程从数据中学习的。
Deep learning is making major advances in solving problems that have resisted the best attempts of the artificial intelligence commu- nity for many years. It has turned out to be very good at discovering intricate structures in high-dimensional data and is therefore applica- ble to many domains of science, business and government. In addition to beating records in image recognition1–4 and speech recognition5–7, it has beaten other machine-learning techniques at predicting the activ- ity of potential drug molecules8, analysing particle accelerator data9,10, reconstructing brain circuits11, and predicting the effects of mutations in non-coding DNA on gene expression and disease12,13. Perhaps more surprisingly, deep learning has produced extremely promising results for various tasks in natural language understanding14, particularly topic classification, sentiment analysis, question answering15 and lan- guage translation16,17.
深度学习在解决多年来一直阻碍人工智能界最佳尝试的问题方面取得了重大进展。事实证明，其非常善于发现高维数据中复杂的结构，因此适用于科学、商业和*的许多领域。除了在图像识别和语音识别方面的记录外，它在预测潜在药物分子的活性、分析粒子加速器数据、重建大脑回路等方面击败了其他机器学习技术，预测非编码DNA突变对基因表达和疾病的影响。更令人惊讶的是，深度学习在自然语言理解的各种任务中产生了非常有希望的结果，尤其是主题分类、情绪分析、问题回答和语言翻译的应用中。
We think that deep learning will have many more successes in the near future because it requires very little engineering by hand, so it can easily take advantage of increases in the amount of available com- putation and data. New learning algorithms and architectures that are currently being developed for deep neural networks will only acceler- ate this progress.
我们相信，在不久的将来，深度学习将取得更多的成就，因为它只需要很少的手工动操作，它可以很容易地从可用计算机和数据量的增加中得益。目前正在为深层神经网络开发的新的学习算法和结构能加速这一进展。

Supervised learning

The most common form of machine learning, deep or not, is super- vised learning. Imagine that we want to build a system that can classify images as containing, say, a house, a car, a person or a pet. We first collect a large data set of images of houses, cars, people and pets, each labelled with its category. During training, the machine is shown an image and produces an output in the form of a vector of scores, one for each category. We want the desired category to have the highest score of all categories, but this is unlikely to happen before training. We compute an objective function that measures the error (or dis- tance) between the output scores and the desired pattern of scores. The machine then modifies its internal adjustable parameters to reduce this error. These adjustable parameters, often called weights, are real numbers that can be seen as ‘knobs’ that define the input–output func- tion of the machine. In a typical deep-learning system, there may be hundreds of millions of these adjustable weights, and hundreds of millions of labelled examples with which to train the machine.
To properly adjust the weight vector, the learning algorithm com- putes a gradient vector that, for each weight, indicates by what amount the error would increase or decrease if the weight were increased by a tiny amount. The weight vector is then adjusted in the opposite direc- tion to the gradient vector.

The objective function, averaged over all the training examples, can be seen as a kind of hilly landscape in the high-dimensional space of weight values. The negative gradient vector indicates the direction of steepest descent in this landscape, taking it closer to a minimum, where the output error is low on average.

In practice, most practitioners use a procedure called stochastic gradient descent (SGD). This consists of showing the input vector for a few examples, computing the outputs and the errors, computing the average gradient for those examples, and adjusting the weights accordingly. The process is repeated for many small sets of examples from the training set until the average of the objective function stops decreasing. It is called stochastic because each small set of examples gives a noisy estimate of the average gradient over all examples. This simple procedure usually finds a good set of weights surprisingly quickly when compared with far more elaborate optimization tech- niques18. After training, the performance of the system is measured on a different set of examples called a test set. This serves to test the generalization ability of the machine — its ability to produce sensible answers on new inputs that it has never seen during training.

Many of the current practical applications of machine learning use linear classifiers on top of hand-engineered features. A two-class linear classifier computes a weighted sum of the feature vector components. If the weighted sum is above a threshold, the input is classified as belonging to a particular category.

Since the 1960s we have known that linear classifiers can only carve their input space into very simple regions, namely half-spaces sepa- rated by a hyperplane19. But problems such as image and speech recog- nition require the input–output function to be insensitive to irrelevant variations of the input, such as variations in position, orientation or illumination of an object, or variations in the pitch or accent of speech, while being very sensitive to particular minute variations (for example, the difference between a white wolf and a breed of wolf-like white dog called a Samoyed). At the pixel level, images of two Samoyeds in different poses and in different environments may be very different from each other, whereas two images of a Samoyed and a wolf in the same position and on similar backgrounds may be very similar to each other. A linear classifier, or any other ‘shallow’ classifier operating on raw pixels could not possibly distinguish the latter two, while putting the former two in the same category.
This is why shallow classifiers require a good feature extractor that solves the selectivity–invariance dilemma — one that produces representations that are selective to the aspects of the image that are important for discrimination, but that are invariant to irrelevant aspects such as the pose of the animal. To make classifiers more powerful, one can use generic non-linear features, as with kernel methods20, but generic features such as those arising with the Gaussian kernel do not allow the learner to general- ize well far from the training examples21. The conventional option is to hand design good feature extractors, which requires a consider- able amount of engineering skill and domain expertise. But this can all be avoided if good features can be learned automatically using a general-purpose learning procedure. This is the key advantage of deep learning.

A deep-learning architecture is a multilayer stack of simple mod- ules, all (or most) of which are subject to learning, and many of which compute non-linear input–output mappings. Each module in the stack transforms its input to increase both the selectivity and the invariance of the representation. With multiple non-linear layers, say a depth of 5 to 20, a system can implement extremely intricate func- tions of its inputs that are simultaneously sensitive to minute details— distinguishing Samoyeds from white wolves — and insensitive to large irrelevant variations such as the background, pose, lighting and surrounding objects.

监督学习

机器学习最常见的形式，无论是深度学习还是非深度学习，都是监督学习。我们可以想象一下，想要建立一个系统，可以将图像分类为包含，比如说，房子、汽车、人或宠物。我们首先收集了一个大数据集的房子，汽车，人和宠物，每一个都有其分类的标签。在训练过程中，机器会显示一个图像，并以分数向量的形式生成输出，每个类别对应一个分数。我们希望理想的类别在所有类别中得分最高，但这不可能在训练前发生。我们计算一个目标函数来衡量输出分数和期望的分数模式之间的误差（或距离）。然后，机器修改其内部可调参数以减少此误差。这些可调参数，通常称为权重，是实数，可以看作是定义机器输入输出功能的“旋钮”。在一个典型的深度学习系统中，可能有数以亿计的可调权重，以及数以亿计的用于训练机器的标签示例。
为了正确调整权重向量，学习算法计算出一个梯度向量，对于每个权重，该向量指示如果权重稍微增加一点，误差会增加或减少多少。然后在与梯度向量相反的方向上调整权重向量。

在所有训练样本上取平均值的目标函数，可以看作是一种高维权重空间中的丘陵景观。负梯度向量表示该景观中最陡下降的方向，使其更接近最小值，输出误差平均较低。

在实践中，大多数实践者使用随机梯度下降（SGD）的程序，这包括显示几个例子的输入向量，计算输出和误差，计算这些例子的平均梯度，并相应地调整权重。从训练集中的许多小样本重复这个过程，直到目标函数的平均值停止下降。之所以称之为随机性，是因为每个小样本集都给出了所有样本的平均梯度的噪声估计。与更精细的优化技术相比，这个简单的过程通常可以很快地找到一组很好的权重。在训练之后，系统的性能将在一组称为测试集的不同示例上进行测量。这有助于测试机器的泛化能力，即它对新输入产生合理答案的能力，而这是它在训练中不会出现。

目前机器学习的许多实际应用是在手工设计的特征之上使用线性分类器。两类线性分类器计算特征向量分量的加权和，如果加权和高于阈值，则输入被归类为属于特定类别。

自20世纪60年代以来，我们就知道线性分类器只能将其输入空间分割成非常简单的区域，即由超平面分离的半空间。但图像和语音识别等问题要求输入-输出函数对输入的不相关变化不敏感，例如对象位置、方向或照明的变化，或音调或口音的变化，同时对特定的细微变化非常敏感（例如，白狼和一种叫萨摩耶的狼样的白狗之间的区别）。在像素级，不同姿势和不同环境下的两个萨摩耶的图像之间可能彼此非常不同，而在相同位置和相似背景下的萨摩耶人和狼的两个图像可能非常相似。在原始像素上操作的线性分类器或任何其他“浅”分类器都不可能区分后两个，而将前两个放在同一个类别中，这就是为什么浅层分类器需要一个很好的特征抽取器来解决选择性-不变性的困境-一个能够产生对图像的某些方面有选择性的表示，而这些方面对于不相关的方面（如动物的姿势）是不变的。为了使分类器更强大，我们可以使用泛型非线性特征，如kernel方法，但泛型特征（如Gaussian kernel产生的特征）不允许学习者在远离训练示例的地方进行泛化。传统的选择是手工设计好的特征提取器，这需要大量的工程技术和领域专业知识。但是，如果可以使用通用的学习过程自动学习好的特征，那么这些都可以避免，这是深度学习的关键优势。

深度学习体系结构是简单模型的多层堆栈，所有（或大部分）模块都要学习，其中许多模块计算非线性输入-输出映射。堆栈中的每个模块转换其输入，以提高表示的选择性和不变性。有了多个非线性层，比如5到20的深度，一个系统可以实现其输入的极其复杂的功能，这些功能同时对微小的细节敏感——区分萨摩耶和白狼——并且对大的不相关的变化（如背景、姿势、光照和周围的物体）不敏感。

图1 |多层神经网络和反传播.a，多层神经网络（由连接点显示）可以扭曲输入空间，使数据类别（其中的例子在红色和蓝色线上）线性可分离。请注意输入空间中的常规网格（显示在左侧）如何通过隐藏单位（显示在中间面板中）进行转换。这是一个仅包含两个输入单元、两个隐藏单元和一个输出单元的示例，但用于对象识别或自然语言处理的网络包含数万或数十万个单元。经C.Olah（http://colah.github.io/）许可http://colah.github.io/转载。b、导数链规则告诉我们如何组成两个小效应（y上的x的小变化和y对z的小变化）。x 中的小变化 =x 首先通过乘以 y/x（即部分导数的定义）在 y 中转换为小变化 μy。同样，变化在 z 中创造了一个变化\z。 将一个方程替换成另一个方程，给出了导数的链规则 — μx 如何通过乘法将 \y/\x 和\z/\x 的乘法转化为\z。当 x、y 和 z 是矢量（导数是雅各比矩阵）时，它也有效。c、用于计算神经网中具有两个隐藏层和一个输出层的前进通道的方程，每个层构成一个模块，通过该模块可以支持反面梯度。在每个层，我们首先计算每个单位的总输入 z，这是下面层中单位输出的加权总和。然后将非线性函数 f（.） 应用于 z 以获得单位的输出。为简单起见，我们省略了偏置词。神经网络中使用的非线性函数包括近年来常用的整流线性单元（ReLU）f（z）=max（0，z）， 以及更传统的西格莫德，如催眠切线，f（z）=（exp（z）=exp（z）/（exp（z）=exp（\z））和物流函数物流，f（z）=1/（1=exp（+z）。d、用于计算后向传递的方程。在每个隐藏层，我们计算与每个单元的输出的误差导数，这是误差导数与上述层中单位的总输入的加权和。然后，我们将与输出的误差导数转换为与输入的误差导数，将其乘以 f（z） 的渐变。在输出层，通过区分成本函数来计算与单位输出的误差导数。如果单位 l 的成本函数为 0.5（yl=tl）2，则这给出了 yl=tl，其中 tl 是目标值。一旦知道 [E/\zkis，下面层中单元 j 连接上权重 wjk 的错误导数只是 yj +E+/zk。

图2 |在卷积网络内。应用于萨摩耶狗图像（左下;和RGB（红色、绿色、蓝色）输入（右下）的图像的典型卷积网络体系结构的每个层的输出（不是滤波器）。每个矩形图像都是与其中一个已学要素的输出对应的要素贴图，在每个图像位置检测到。信息自下而上流动，较低级别的要素充当定向边缘探测器，并计算输出中每个图像类的分数。ReLU，整流线性单元。

Backpropagation to train multilayer architectures

From the earliest days of pattern recognition, the aim of researchers has been to replace hand-engineered features with trainable multilayer networks, but despite its simplicity, the solution was not widely understood until the mid 1980s. As it turns out, multilayer architectures can be trained by simple stochastic gradient descent. As long as the modules are relatively smooth functions of their inputs and of their internal weights, one can compute gradients using the backpropagation procedure. The idea that this could be done, and that it worked, was discovered independently by several different groups during the 1970s and 1980s.

The backpropagation procedure to compute the gradient of an objective function with respect to the weights of a multilayer stack of modules is nothing more than a practical application of the chain rule for derivatives. The key insight is that the derivative (or gradi- ent) of the objective with respect to the input of a module can be computed by working backwards from the gradient with respect to the output of that module (or the input of the subsequent module) (Fig. 1). The backpropagation equation can be applied repeatedly to propagate gradients through all modules, starting from the output at the top (where the network produces its prediction) all the way to the bottom (where the external input is fed). Once these gradients have been computed, it is straightforward to compute the gradients with respect to the weights of each module.

Many applications of deep learning use feedforward neural net- work architectures (Fig. 1), which learn to map a fixed-size input (for example, an image) to a fixed-size output (for example, a prob- ability for each of several categories). To go from one layer to the next, a set of units compute a weighted sum of their inputs from the previous layer and pass the result through a non-linear function. At present, the most popular non-linear function is the rectified linear unit (ReLU), which is simply the half-wave rectifier f(z) = max(z, 0). In past decades, neural nets used smoother non-linearities, such as tanh(z) or 1/(1 + exp(−z)), but the ReLU typically learns much faster in networks with many layers, allowing training of a deep supervised network without unsupervised pre-training28. Units that are not in the input or output layer are conventionally called hidden units. The hidden layers can be seen as distorting the input in a non-linear way so that categories become linearly separable by the last layer (Fig. 1).

In the late 1990s, neural nets and backpropagation were largely forsaken by the machine-learning community and ignored by the computer-vision and speech-recognition communities. It was widely thought that learning useful, multistage, feature extractors with little prior knowledge was infeasible. In particular, it was commonly thought that simple gradient descent would get trapped in poor local minima — weight configurations for which no small change would reduce the average error.

In practice, poor local minima are rarely a problem with large net- works. Regardless of the initial conditions, the system nearly always reaches solutions of very similar quality. Recent theoretical and empirical results strongly suggest that local minima are not a serious issue in general. Instead, the landscape is packed with a combinato- rially large number of saddle points where the gradient is zero, and the surface curves up in most dimensions and curves down in the remainder29,30. The analysis seems to show that saddle points with only a few downward curving directions are present in very large numbers, but almost all of them have very similar values of the objec-tive function. Hence, it does not much matter which of these saddle points the algorithm gets stuck at.

Interest in deep feedforward networks was revived around 2006 (refs 31–34) by a group of researchers brought together by the Cana- dian Institute for Advanced Research (CIFAR). The researchers intro- duced unsupervised learning procedures that could create layers of feature detectors without requiring labelled data. The objective in learning each layer of feature detectors was to be able to reconstruct or model the activities of feature detectors (or raw inputs) in the layer below. By ‘pre-training’ several layers of progressively more complex feature detectors using this reconstruction objective, the weights of a deep network could be initialized to sensible values. A final layer of output units could then be added to the top of the network and the whole deep system could be fine-tuned using standard backpropaga- tion33–35. This worked remarkably well for recognizing handwritten digits or for detecting pedestrians, especially when the amount of labelled data was very limited36.

The first major application of this pre-training approach was in speech recognition, and it was made possible by the advent of fast graphics processing units (GPUs) that were convenient to program37 and allowed researchers to train networks 10 or 20 times faster. In 2009, the approach was used to map short temporal windows of coef- ficients extracted from a sound wave to a set of probabilities for the various fragments of speech that might be represented by the frame in the centre of the window. It achieved record-breaking results on a standard speech recognition benchmark that used a small vocabu- lary38 and was quickly developed to give record-breaking results on a large vocabulary task39. By 2012, versions of the deep net from 2009 were being developed by many of the major speech groups6 and were already being deployed in Android phones. For smaller data sets, unsupervised pre-training helps to prevent overfitting40, leading to significantly better generalization when the number of labelled exam- ples is small, or in a transfer setting where we have lots of examples for some ‘source’ tasks but very few for some ‘target’ tasks. Once deep learning had been rehabilitated, it turned out that the pre-training stage was only needed for small data sets.

There was, however, one particular type of deep, feedforward network that was much easier to train and generalized much better than networks with full connectivity between adjacent layers. This was the convolutional neural network (ConvNet)41,42. It achieved many practical successes during the period when neural networks were out of favour and it has recently been widely adopted by the computer- vision community.

反向传播训练多层体系结构

从最早的模式识别，研究者的目的一直在用可训练的多层网络来代替手工设计的特征，但是尽管它很简单，直到20世纪80年代中期，这个解决方案才被广泛理解。事实证明，多层结构可以通过简单的随机梯度下降来训练。只要模块是其输入和内部权重的相对平滑函数，就可以使用反向传播程序计算梯度。在20世纪70年代和80年代，几个不同的团体独立地发现了这个想法，即这是可以做到的，而且是有效的。

用反向传播法计算一个目标函数相对于一个多层模块栈的权重的梯度，不过是导数链规则的一个实际应用。关键的见解是，目标相对于模块输入的导数（或梯度）可以通过从相对于该模块输出（或后续模块的输入）的梯度向后计算（图1）。反向传播方程可重复应用于在所有模块中传播梯度，从顶部的输出（网络产生其预测）一直到底部（外部输入被馈送）。一旦计算了这些梯度，就很容易计算出相对于每个模块权重的梯度。

深度学习的许多应用使用前馈神经网络结构（图1），它学习将固定大小的输入（例如，图像）映射到固定大小的输出（例如，几个类别中的每一个的概率）。为了从一层到下一层，一组单元计算上一层输入的加权和，并将结果传递给一个非线性函数。目前最流行的非线性函数是整流线性单元（ReLU），即半波整流器f（z）=max（z，0）。在过去的几十年里，神经网络使用更平滑的非线性，例如tanh（z）或1/（1+exp（−z）），但ReLU通常在多层网络中学习得更快，允许在无监督预训练的情况下训练深度监督网络28。不在输入或输出层的单元通常称为隐藏单元。隐藏层可以被视为以非线性方式扭曲输入，使得类别可以由最后一层线性分离（图1）。

在20世纪90年代末，神经网络和反向传播在很大程度上被机器学习界所抛弃，而被计算机视觉和语音识别界所忽视。人们普遍认为，学习有用的、多阶段的、具有少量先验知识的特征抽取器是不可行的。特别是，大家普遍认为简单的梯度下降会陷入局部最小权配置中，对于这种配置，任何小的变化都会降低平均误差。

在实践中，局部极小值很少是大型网络的问题。不管初始条件如何，系统几乎总能得到质量非常相似的解。最近的理论和实证结果强烈表明，局部极小值一般不是一个严重的问题。取而代之的是，景观中有大量的鞍点，坡度为零，而曲面在大多数维度上呈上升趋势，而在剩余的维度上则向下弯曲。分析表明，只有几个向下弯曲方向的鞍点数量非常多，但几乎所有鞍点的目标函数值都非常相似，因此，算法在这些鞍点中哪一个被卡住并不重要。

加拿大高级研究所（CIFAR）召集的一组研究人员在2006年左右恢复了对深层前馈网络的兴趣。研究人员引入了无监督的学习程序，这些程序可以创建特征检测器层，而无需标记数据。学习特征检测器每一层的目的是能够在下一层中重建或建模特征检测器（或原始输入）的活动。通过使用此重建目标对多层逐渐复杂的特征检测器进行“预训练”，可以将深度网络的权重初始化为合理的值。然后可以将输出单元的最后一层添加到网络的顶部，并且可以使用标准反向传播对整个深度系统进行微调。这对于识别手写数字或检测行人非常有效，特别是在标记数据量非常有限的情况下。

这种预训练方法的第一个主要应用是语音识别，而快速图形处理单元（GPU）的出现使编程成为可能，并且使研究人员训练网络的速度提高了10或20倍，从而使之成为可能。2009年，该方法用于将从声波中提取的系数的短时窗映射到各种语音片段的概率集，这些概率可能由窗口中心的帧表示。它在使用少量词汇的标准语音识别基准上达到了破纪录的结果，并迅速开发以在大型词汇任务上提供破纪录的结果。到2012年，许多主要的语音小组正在开发2009年以来的深网版本6，并且已经在Android手机中进行了部署。对于较小的数据集，无监督的预训练有助于防止过度拟合，从而在标记的样本数量较小时或在转移设置中，我们有很多“源”任务的例子很多，但很少的情况下，泛化效果显着提高一些“目标”任务。恢复了深度学习后，事实证明，仅对于小型数据集才需要进行预训练。

但是，存在一种特定类型的深层前馈网络，它比相邻层之间具有完全连接的网络更容易训练和推广，这就是卷积神经网络（ConvNet）。在神经网络使用下降期间，它取得了许多实际的成功，并且最近被计算机视觉界广泛采用。

Convolutional neural networks

ConvNets are designed to process data that come in the form of multiple arrays, for example a colour image composed of three 2D arrays containing pixel intensities in the three colour channels. Many data modalities are in the form of multiple arrays: 1D for signals and sequences, including language; 2D for images or audio spectrograms; and 3D for video or volumetric images. There are four key ideas behind ConvNets that take advantage of the properties of natural signals: local connections, shared weights, pooling and the use of many layers.

The architecture of a typical ConvNet (Fig. 2) is structured as a series of stages. The first few stages are composed of two types of layers: convolutional layers and pooling layers. Units in a convolu- tional layer are organized in feature maps, within which each unit is connected to local patches in the feature maps of the previous layer through a set of weights called a filter bank. The result of this local weighted sum is then passed through a non-linearity such as a ReLU. All units in a feature map share the same filter bank. Differ- ent feature maps in a layer use different filter banks. The reason for this architecture is twofold. First, in array data such as images, local groups of values are often highly correlated, forming distinctive local motifs that are easily detected. Second, the local statistics of images and other signals are invariant to location. In other words, if a motif can appear in one part of the image, it could appear anywhere, hence the idea of units at different locations sharing the same weights and detecting the same pattern in different parts of the array. Mathemati- cally, the filtering operation performed by a feature map is a discrete convolution, hence the name.

Although the role of the convolutional layer is to detect local con- junctions of features from the previous layer, the role of the pooling layer is to merge semantically similar features into one. Because the relative positions of the features forming a motif can vary somewhat, reliably detecting the motif can be done by coarse-graining the posi- tion of each feature. A typical pooling unit computes the maximum of a local patch of units in one feature map (or in a few feature maps). Neighbouring pooling units take input from patches that are shifted by more than one row or column, thereby reducing the dimension of the representation and creating an invariance to small shifts and dis- tortions. Two or three stages of convolution, non-linearity and pool- ing are stacked, followed by more convolutional and fully-connected layers. Backpropagating gradients through a ConvNet is as simple as through a regular deep network, allowing all the weights in all the filter banks to be trained.

Deep neural networks exploit the property that many natural sig-nals are compositional hierarchies, in which higher-level features are obtained by composing lower-level ones. In images, local combinations of edges form motifs, motifs assemble into parts, and parts form objects. Similar hierarchies exist in speech and text from sounds to phones, phonemes, syllables, words and sentences. The pooling allows representations to vary very little when elements in the previous layer vary in position and appearance.

The convolutional and pooling layers in ConvNets are directly inspired by the classic notions of simple cells and complex cells in visual neuroscience43, and the overall architecture is reminiscent of the LGN–V1–V2–V4–IT hierarchy in the visual cortex ventral path- way44. When ConvNet models and monkeys are shown the same pic- ture, the activations of high-level units in the ConvNet explains half of the variance of random sets of 160 neurons in the monkey’s infer- otemporal cortex45. ConvNets have their roots in the neocognitron46, the architecture of which was somewhat similar, but did not have an end-to-end supervised-learning algorithm such as backpropagation. A primitive 1D ConvNet called a time-delay neural net was used for the recognition of phonemes and simple words.

There have been numerous applications of convolutional net- works going back to the early 1990s, starting with time-delay neu- ral networks for speech recognition47 and document reading42. The document reading system used a ConvNet trained jointly with a probabilistic model that implemented language constraints. By the late 1990s this system was reading over 10% of all the cheques in the United States. A number of ConvNet-based optical character recog- nition and handwriting recognition systems were later deployed by Microsoft49. ConvNets were also experimented with in the early 1990s for object detection in natural images, including faces and hands50,51, and for face recognition.

卷积神经网络

ConvNets被设计为处理以多个阵列形式出现的数据，例如由三个二维通道组成的彩色图像，其中三个二维通道在三个彩色通道中包含像素强度。许多数据形式以多个数组的形式出现：一维用于信号和序列，包括语言；2D用于图像或音频频谱图和3D视频或立体图像。ConvNets有四个利用自然信号属性的关键思想：本地连接，共享权重，池化和多层使用。

典型的ConvNet的体系结构由一系列阶段构成。前几个阶段由两种类型的层组成：卷积层和池化层。卷积层中的单元组织在特征图中，其中每个单元通过一组称为过滤器的权重连接到上一层特征图中的局部面片。然后，该局部加权和的结果将通过非线性（例如ReLU）传递。特征图中的所有单位共享相同的过滤器组。图层中的不同要素图使用不同的滤镜库。这种架构的原因有两个。首先，在诸如图像的阵列数据中，值的局部组通常高度相关，形成易于检测的独特局部图案。其次，图像和其他信号的局部统计量对于位置是不变的。换句话说，如果图案可以出现在图像的一部分中，则它可以出现在任何地方，因此，位于不同位置的单元在阵列的不同部分共享相同的权重并检测相同的图案的想法。从数学上讲，由特征图执行的过滤操作是离散卷积，故其因此得名。

尽管卷积层的作用是检测上一层的要素的局部结合，但池化层的作用是将语义相似的要素合并为一个。由于形成图案的特征的相对位置可能会有所不同，因此可以通过对每个特征的位置进行粗粒度来可靠地检测出图案。典型的池化单元计算一个要素地图（或几个要素地图）中局部补丁的最大值。相邻的合并单元从位移了多行或多列的面片中获取输入，从而减小了表示的尺寸，并产生了对较小的位移和失真的不变性。卷积，非线性和合并的两个或三个阶段被堆叠，然后是更多卷积和完全连接的层。通过ConvNet进行反向传播的梯度与通过常规深度网络一样简单，从而可以训练所有滤波器组中的所有权重。

深度神经网络利用了许多自然信号是组成层次结构的特性，其中高层次的特征是通过组合低层次的特征来获得的。 在图像中，边缘的局部组合形成图案，图案组合成零件，而零件形成对象。从声音到电话、音素、音节、单词和句子、语音和文本中也存在类似的层次结构。当上一层中的元素的位置和外观发生变化时，合并使表示的变化很小。

ConvNets中的卷积和池化层直接受到视觉神经科学中简单细胞和复杂细胞的经典概念的启发，整个架构让人联想到视觉皮层腹侧通路中的LGN–V1–V2–V4–IT层次结构。当ConvNet模型和猴子显示相同的图片时，ConvNet中高层单元的**可以解释猴子下颞叶皮层中160个神经元的随机集合的一半变化。ConvNets的根源是neocognitron，其架构有些相似，但没有反向传播等端到端监督学习算法，使用称为时延神经网络的原始一维ConvNet来识别音素和简单单词。

卷积网络的应用可以追溯到1990年代初，首先是用于语音识别和文档读取的时延神经网络。 该文档阅读系统使用了一个ConvNet，并与一个实现语言约束的概率模型一起进行了培训。到1990年代后期，该系统已读取了美国所有支票的10％以上。后来，Microsoft部署了许多基于ConvNet的光学字符识别和手写识别系统,在1990年代初期，还对ConvNets进行了试验，以检测自然图像中的物体，包括面部和手部，以及面部识别。

图3 |从图像到文本。由循环神经网络 （RNN） 生成的字幕，作为额外的输入，从测试图像中提取由深度卷积神经网络 （CNN） 提取的表示形式，RNN 训练将图像的高级别表示"翻译"到字幕（顶部）。经参考文献102许可转载。当 RNN 能够将注意力集中在输入图像的不同位置（中间和底部;较轻的补丁被给予更多关注）时，我们发现 86 它利用这一点将图像更好地"翻译"成字幕。

Image understanding with deep convolutional networks

Since the early 2000s, ConvNets have been applied with great success to the detection, segmentation and recognition of objects and regions in images. These were all tasks in which labelled data was relatively abun- dant, such as traffic sign recognition53, the segmentation of biological images54 particularly for connectomics55, and the detection of faces, text, pedestrians and human bodies in natural images36,50,51,56–58. A major recent practical success of ConvNets is face recognition59.

Importantly, images can be labelled at the pixel level, which will have applications in technology, including autonomous mobile robots and self-driving cars60,61. Companies such as Mobileye and NVIDIA are using such ConvNet-based methods in their upcoming vision sys- tems for cars. Other applications gaining importance involve natural language understanding14 and speech recognition.

Despite these successes, ConvNets were largely forsaken by the mainstream computer-vision and machine-learning communities until the ImageNet competition in 2012. When deep convolutional networks were applied to a data set of about a million images from the web that contained 1,000 different classes, they achieved spec- tacular results, almost halving the error rates of the best compet- ing approaches1. This success came from the efficient use of GPUs, ReLUs, a new regularization technique called dropout62, and tech- niques to generate more training examples by deforming the existing ones. This success has brought about a revolution in computer vision; ConvNets are now the dominant approach for almost all recognition and detection tasks4,58,59,63–65 and approach human performance on some tasks. A recent stunning demonstration combines ConvNets and recurrent net modules for the generation of image captions (Fig. 3).

Recent ConvNet architectures have 10 to 20 layers of ReLUs, hun- dreds of millions of weights, and billions of connections between units. Whereas training such large networks could have taken weeks only two years ago, progress in hardware, software and algorithm parallelization have reduced training times to a few hours.

The performance of ConvNet-based vision systems has caused most major technology companies, including Google, Facebook, Microsoft, IBM, Yahoo!, Twitter and Adobe, as well as a quickly growing number of start-ups to initiate research and development projects and to deploy ConvNet-based image understanding products and services.

ConvNets are easily amenable to efficient hardware implemen- tations in chips or field-programmable gate arrays66,67. A number of companies such as NVIDIA, Mobileye, Intel, Qualcomm and Samsung are developing ConvNet chips to enable real-time vision applications in smartphones, cameras, robots and self-driving cars.

利用深度卷积网络进行图像理解

自2000年代初以来，ConvNets已成功应用于检测，分割和识别图像中的对象和区域。这些都是带有标记数据的相对丰富的任务，例如交通标志识别，生物图像分割，尤其是用于连接组学，以及在自然图像中检测人脸、文字、行人和人体，ConvNets最近的一项重大实践成功是人脸识别。

重要的是，可以在像素级别标记图像，这将在技术中得到应用，包括自动驾驶机器人和自动驾驶汽车。 诸如Mobileye和NVIDIA之类的公司正在其即将推出的汽车视觉系统中使用这种基于ConvNet的方法，其他日益重要的应用包括自然语言理解和语音识别。

尽管取得了这些成功，但ConvNet在很大程度上被主流的计算机视觉和机器学习社区所抛弃，直到2012年ImageNet竞赛。当深度卷积网络应用于来自网络的大约一百万个图像的数据集时，其中包含1,000个不同的类别， 他们取得了惊人的成绩，几乎使最佳竞争方法的错误率降低了一半。 成功的原因是有效利用了GPU，ReLU，一种称为dropout62的新正则化技术，以及通过使现有示例变形而生成更多训练示例的技术。 这一成功带来了计算机视觉的一场革命。 现在，ConvNets是几乎所有识别和检测任务的主要方法4、58、59、63-65，并在某些任务上采用人类绩效。 最近的惊人演示结合了ConvNets和递归网络模块，用于生成图像字幕（图3）。

最新的ConvNet架构具有10到20层ReLU，数以亿计的重量以及单元之间的数十亿个连接。尽管培训如此大型的网络可能仅在两年前才花了几周的时间，但是硬件，软件和算法并行化方面的进步已将培训时间减少到几个小时。

基于ConvNet的视觉系统的性能已引起大多数主要技术公司的发展，其中包括Google，Facebook，Microsoft，IBM，Yahoo！，Twitter和Adobe，以及数量迅速增长的初创公司启动了研发项目，部署基于ConvNet的图像理解产品和服务。

ConvNets易于适应芯片或现场可编程门阵列中的高效硬件实现66,67，NVIDIA，Mobileye，英特尔，高通和三星等多家公司正在开发ConvNet芯片，以支持智能手机，相机，机器人和自动驾驶汽车中的实时视觉应用。

Distributed representations and language processing

Deep-learning theory shows that deep nets have two different expo- nential advantages over classic learning algorithms that do not use distributed representations21. Both of these advantages arise from the power of composition and depend on the underlying data-generating distribution having an appropriate componential structure40. First, learning distributed representations enable generalization to new combinations of the values of learned features beyond those seen during training (for example, 2n combinations are possible with n binary features)68,69. Second, composing layers of representation in a deep net brings the potential for another exponential advantage70 (exponential in the depth).

The hidden layers of a multilayer neural network learn to repre- sent the network’s inputs in a way that makes it easy to predict the target outputs. This is nicely demonstrated by training a multilayer neural network to predict the next word in a sequence from a local context of earlier words71. Each word in the context is presented to the network as a one-of-N vector, that is, one component has a value of 1 and the rest are 0. In the first layer, each word creates a different pattern of activations, or word vectors (Fig. 4). In a language model, the other layers of the network learn to convert the input word vec- tors into an output word vector for the predicted next word, which can be used to predict the probability for any word in the vocabulary to appear as the next word. The network learns word vectors that contain many active components each of which can be interpreted as a separate feature of the word, as was first demonstrated27 in the context of learning distributed representations for symbols. These semantic features were not explicitly present in the input. They were discovered by the learning procedure as a good way of factorizing the structured relationships between the input and output symbols into multiple ‘micro-rules’. Learning word vectors turned out to also work very well when the word sequences come from a large corpus of real text and the individual micro-rules are unreliable71. When trained to predict the next word in a news story, for example, the learned word vectors for Tuesday and Wednesday are very similar, as are the word vectors for Sweden and Norway. Such representations are called distributed representations because their elements (the features) are not mutually exclusive and their many configurations correspond to the variations seen in the observed data. These word vectors are composed of learned features that were not determined ahead of time by experts, but automatically discovered by the neural network. Vector representations of words learned from text are now very widely used in natural language applications14,17,72–76.

The issue of representation lies at the heart of the debate between the logic-inspired and the neural-network-inspired paradigms for cognition. In the logic-inspired paradigm, an instance of a symbol is something for which the only property is that it is either identical or non-identical to other symbol instances. It has no internal structure that is relevant to its use; and to reason with symbols, they must be bound to the variables in judiciously chosen rules of inference. By contrast, neural networks just use big activity vectors, big weight matrices and scalar non-linearities to perform the type of fast ‘intui- tive’ inference that underpins effortless commonsense reasoning.

Before the introduction of neural language models71, the standard approach to statistical modelling of language did not exploit distrib- uted representations: it was based on counting frequencies of occur- rences of short symbol sequences of length up to N (called N-grams). The number of possible N-grams is on the order of VN, where V is the vocabulary size, so taking into account a context of more than a handful of words would require very large training corpora. N-grams treat each word as an atomic unit, so they cannot generalize across semantically related sequences of words, whereas neural language models can because they associate each word with a vector of real valued features, and semantically related words end up close to each other in that vector space (Fig. 4).

分布式表示和语言处理

深度学习理论表明，与不使用分布式表示的经典学习算法相比，深网具有两个不同的指数优势。 这两个优点都来自于组合的力量，并取决于具有适当组件结构的底层数据生成分布。首先，学习分布式表示可以将学习到的特征值的新组合推广到训练期间看不到的那些新组合（例如，使用n个二进制特征可以进行2n个组合）。 其次，在一个深层网络中构成表示层会带来另一个指数优势（深度指数）。

多层神经网络的隐藏层学习以易于预测目标输出的方式来表示网络的输入。通过训练多层神经网络从较早单词的局部上下文中预测序列中的下一个单词，可以很好地证明这一点。上下文中的每个单词都以N个向量的形式呈现给网络，也就是说，一个组成部分的值为1，其余均为0。在第一层中，每个单词都会创建不同的**模式，或者单词向量（图4）。在语言模型中，网络的其他层学习将输入的词向量转换为预测的下一个词的输出词向量，该向量可用于预测词汇表中任何词出现为下一个词的可能性字。网络学习包含许多有效成分的单词向量，每个成分都可以解释为单词的一个独立特征，如在学习符号的分布式表示的背景下首次证明的那样。这些语义特征未在输入中明确显示。通过学习过程可以发现它们，这是将输入和输出符号之间的结构化关系分解为多个“微规则”的好方法。当单词序列来自大量的真实文本并且单个微规则不可靠时，学习单词向量也可以很好地工作。例如，在接受培训以预测新闻故事中的下一个单词时，周二和周三学到的单词向量与瑞典和挪威的单词向量非常相似。这样的表示称为分布式表示，因为它们的元素（特征）不是互斥的，并且它们的许多配置对应于在观察到的数据中看到的变化。这些词向量由专家事先未确定但由神经网络自动发现的学习特征组成。从文本中学到的单词的矢量表示形式现在在自然语言应用中得到了广泛的应用。

表征问题是逻辑启发和神经网络启发的认知范式之间争论的核心。 在逻辑启发范式中，符号实例是某些事物，其唯一属性是它与其他符号实例相同或不同。 它没有与其使用相关的内部结构； 为了用符号推理，必须将它们绑定到明智选择的推理规则中的变量。 相比之下，神经网络仅使用较大的活动矢量，较大的权重矩阵和标量非线性来执行快速的“直观”推理，从而支撑了轻松的常识推理。

在引入神经语言模型之前，语言统计建模的标准方法并未利用分布式表示形式：它是基于对长度不超过N的短符号序列（称为N-gram）的出现频率进行计数。 可能的N元语法的数量在VN的数量级上，其中V是词汇量，因此考虑到少数单词的上下文，将需要非常大的训练语料库。N-gram将每个单词视为一个原子单元，因此它们无法概括语义相关的单词序列，而神经语言模型则可以将它们与实值特征向量关联在一起，而语义相关的单词最终彼此靠近在该向量空间中（图4）。

图4 |可视化学习的单词矢量。左侧是为建模语言所学的单词表示的插图，使用 t-SNE 算法 103 将非线性投影到 2D 进行可视化。右侧是英语到法语编码器+解码器循环神经网络所学短语的二D表示形式。可以观察到，语义上相似的单词或单词序列映射到附近的表示形式。单词的分布式表示是通过使用回发法来共同学习每个单词的表示形式和预测目标数量的函数（如序列中的下一个单词（用于语言建模）或整个翻译单词序列（用于机器翻译）。

图 5 |循环神经网络及其前向计算中涉及的计算时间展开。人工神经元（例如，在节点 s 下分组的隐藏单元，值 st 在时间 t）在以前的时间步长中从其他神经元获取输入（这用黑色正方形表示，表示左侧一个时间步长的延迟）。这样，循环神经网络就可以将具有元素 xt 的输入序列映射到具有 ot 元素的输出序列中，每个输入序列都依赖于所有xtʹ（对于 tʹ\t）。在每个时间步数上使用相同的参数（矩阵 U、V、W）。许多其他体系结构是可能的，包括网络可以生成一系列输出（例如单词）的变体，每个输出都用作下一个时间步骤的输入。反传播算法（图1）可以直接应用于右侧展开的网络的计算图形，以计算针对所有状态和所有参数的总误差（例如，生成正确输出序列的日志概率）的导数。

Recurrent neural networks

When backpropagation was first introduced, its most exciting use was for training recurrent neural networks (RNNs). For tasks that involve sequential inputs, such as speech and language, it is often better to use RNNs (Fig. 5). RNNs process an input sequence one element at a time, maintaining in their hidden units a ‘state vector’ that implicitly contains information about the history of all the past elements of the sequence. When we consider the outputs of the hidden units at different discrete time steps as if they were the outputs of different neurons in a deep multilayer network (Fig. 5, right), it becomes clear how we can apply backpropagation to train RNNs.

RNNs are very powerful dynamic systems, but training them has proved to be problematic because the backpropagated gradients either grow or shrink at each time step, so over many time steps they typically explode or vanish.

Thanks to advances in their architecture79,80 and ways of training them81,82, RNNs have been found to be very good at predicting the next character in the text83 or the next word in a sequence75, but they can also be used for more complex tasks. For example, after reading an English sentence one word at a time, an English ‘encoder’ network can be trained so that the final state vector of its hidden units is a good representation of the thought expressed by the sentence. This thought vector can then be used as the initial hidden state of (or as extra input to) a jointly trained French ‘decoder’ network, which outputs a prob- ability distribution for the first word of the French translation. If a particular first word is chosen from this distribution and provided as input to the decoder network it will then output a probability dis- tribution for the second word of the translation and so on until a full stop is chosen17,72,76. Overall, this process generates sequences of French words according to a probability distribution that depends on the English sentence. This rather naive way of performing machine translation has quickly become competitive with the state-of-the-art, and this raises serious doubts about whether understanding a sen- tence requires anything like the internal symbolic expressions that are manipulated by using inference rules. It is more compatible with the view that everyday reasoning involves many simultaneous analogies that each contribute plausibility to a conclusion.

Instead of translating the meaning of a French sentence into an English sentence, one can learn to ‘translate’ the meaning of an image into an English sentence (Fig. 3). The encoder here is a deep Con- vNet that converts the pixels into an activity vector in its last hidden layer. The decoder is an RNN similar to the ones used for machine translation and neural language modelling. There has been a surge of interest in such systems recently (see examples mentioned in ref. 86). RNNs, once unfolded in time (Fig. 5), can be seen as very deep feedforward networks in which all the layers share the same weights. Although their main purpose is to learn long-term dependencies, theoretical and empirical evidence shows that it is difficult to learn to store information for very long78.

To correct for that, one idea is to augment the network with an explicit memory. The first proposal of this kind is the long short-term memory (LSTM) networks that use special hidden units, the natural behaviour of which is to remember inputs for a long time79. A special unit called the memory cell acts like an accumulator or a gated leaky neuron: it has a connection to itself at the next time step that has a weight of one, so it copies its own real-valued state and accumulates the external signal, but this self-connection is multiplicatively gated by another unit that learns to decide when to clear the content of the memory.
LSTM networks have subsequently proved to be more effective than conventional RNNs, especially when they have several layers for each time step87, enabling an entire speech recognition system that goes all the way from acoustics to the sequence of characters in the transcription. LSTM networks or related forms of gated units are also currently used for the encoder and decoder networks that perform so well at machine translation17,72,76.

Over the past year, several authors have made different proposals to augment RNNs with a memory module. Proposals include the Neural Turing Machine in which the network is augmented by a ‘tape-like’ memory that the RNN can choose to read from or write to88, and memory networks, in which a regular network is augmented by a kind of associative memory89. Memory networks have yielded excel- lent performance on standard question-answering benchmarks. The memory is used to remember the story about which the network is later asked to answer questions

Beyond simple memorization, neural Turing machines and mem-ory networks are being used for tasks that would normally require reasoning and symbol manipulation. Neural Turing machines can be taught ‘algorithms’. Among other things, they can learn to output a sorted list of symbols when their input consists of an unsorted sequence in which each symbol is accompanied by a real value that indicates its priority in the list88. Memory networks can be trained to keep track of the state of the world in a setting similar to a text adventure game and after reading a story, they can answer questions that require complex inference90. In one test example, the network is shown a 15-sentence version of the The Lord of the Rings and correctly answers questions such as “where is Frodo now?”

循环神经网络

首次引入反向传播时，其最令人振奋的用途是训练循环神经网络（RNN）。对于涉及顺序输入的任务，例如语音和语言，通常最好使用RNN（图5）。 RNN一次处理一个输入序列一个元素，并在其隐藏单元中维护一个“状态向量”，该向量隐式包含有关该序列的所有过去元素的历史信息。 当我们将隐藏单元在不同离散时间步的输出视为是深层多层网络中不同神经元的输出时（图5，右），很清楚地知道我们如何应用反向传播来训练RNN。

RNN是非常强大的动态系统，但事实证明，训练它们是有问题的，因为反向传播的梯度在每个时间步长都会增大或缩小，因此在许多时间步长上它们通常会爆炸或消失。

得益于其体系结构79,80的进步以及对其进行训练的方法，人们发现RNN非常擅长预测文本中的下一个字符83或序列中的下一个单词75，但它们也可以用于更复杂的任务。例如，一次阅读一个英语句子后，可以训练一个英语“编码器”网络，以使其隐藏单元的最终状态向量很好地表示该句子表达的思想。然后，可以将此思想向量用作联合训练的法语“解码器”网络的初始隐藏状态（或作为其额外输入），该网络将输出法语翻译的第一个单词的概率分布。如果从该分布中选择了一个特定的第一个单词并将其作为输入提供给解码器网络，则它将输出翻译的第二个单词的概率分布，依此类推，直到选择了句号。总体而言，此过程根据取决于英语句子的概率分布生成法语单词序列。这种相当幼稚的执行机器翻译的方式已迅速与最新技术竞争，这使人们对是否要理解语义是否需要诸如通过使用推理规则进行操纵的内部符号表达式等问题产生了严重的怀疑。日常推理涉及许多同时进行的类比，每个类比都有助于得出结论，这更符合以下观点。

与其将法语句子的含义翻译成英语句子，不如学习将图像的含义“翻译”成英语句子（图3）。 这里的编码器是一个深层的ConNet，它将像素转换为最后一个隐藏层中的活动矢量。 解码器是一种RNN，类似于用于机器翻译和神经语言建模的RNN。 最近，对这种系统的兴趣激增（参见参考文献86中提到的示例），RNN一旦及时展开（图5），就可以视为非常深的前馈网络，其中所有层共享相同的权重。尽管它们的主要目的是学习长期依赖关系，但理论和经验证据表明，很难学习将信息存储很长时间。

为了解决这个问题，一个想法是用显式内存扩展网络。此类第一个建议是使用特殊隐藏单元的长短期记忆（LSTM）网络，其自然行为是长时间记住输入。称为存储单元的特殊单元的作用类似于累加器或门控泄漏神经元：它在下一时间步与其自身具有连接，其权重为1，因此它复制其自己的实值状态并累积外部信号，但是这种自我连接是由另一个单元乘法控制的，该单元学会决定何时清除存储器的内容。
随后证明LSTM网络比常规RNN更有效，特别是当它们在每个时间步都有多层时，使整个语音识别系统从声学到转录中的字符序列都一路走来。LSTM网络或相关形式的门控单元目前也用于编码器和解码器网络，它们在机器翻译方面表现出色。

在过去的一年中，几位作者提出了不同的建议，以使用内存模块扩展RNN。 提议包括神经图灵机，其中网络由RNN可以选择读取或写入的“类磁带”存储器来增强，以及存储网络，其中常规网络由一种关联存储器来增强，从而使网络得以扩展。 内存网络在标准问答基准上表现出卓越的性能。 内存用于记住故事，有关该故事后来被要求网络回答问题。

除了简单的记忆外，神经网络图灵机和内存网络还用于通常需要推理和符号操作的任务。 神经图灵机可以被称为“算法”。 除其他事项外，当他们的输入由未排序的序列组成时，他们可以学会输出已排序的符号列表，其中每个符号都带有一个实数值，该实数值指示其在列表中的优先级88。 可以训练记忆网络，使其在类似于文字冒险游戏的环境中跟踪世界状况，阅读故事后，它们可以回答需要复杂推理的问题。 在一个测试示例中，该网络显示了15句的《指环王》，并正确回答了诸如“ Frodo现在在哪里？”之类的问题。

The future of deep learning

Unsupervised learning91–98 had a catalytic effect in reviving interest in deep learning, but has since been overshadowed by the successes of purely supervised learning. Although we have not focused on it in this Review, we expect unsupervised learning to become far more important in the longer term. Human and animal learning is largely unsupervised: we discover the structure of the world by observing it, not by being told the name of every object.
Human vision is an active process that sequentially samples the optic array in an intelligent, task-specific way using a small, high-resolution fovea with a large, low-resolution surround. We expect much of the future progress in vision to come from systems that are trained end-to- end and combine ConvNets with RNNs that use reinforcement learning to decide where to look. Systems combining deep learning and rein- forcement learning are in their infancy, but they already outperform passive vision systems99 at classification tasks and produce impressive results in learning to play many different video games100.

Natural language understanding is another area in which deep learn-ing is poised to make a large impact over the next few years. We expect systems that use RNNs to understand sentences or whole documents will become much better when they learn strategies for selectively attending to one part at a time76,86.
Ultimately, major progress in artificial intelligence will come about through systems that combine representation learning with complex reasoning. Although deep learning and simple reasoning have been used for speech and handwriting recognition for a long time, new paradigms are needed to replace rule-based manipulation of symbolic expressions by operations on large vectors101

深度学习的未来

无监督学习在重新激发人们对深度学习的兴趣方面起到了催化作用，但此后被纯粹的监督学习的成功所掩盖。尽管我们在本评论中并未对此进行关注，但是我们希望从长远来看，无监督学习将变得越来越重要。人类和动物的学习在很大程度上不受监督：我们通过观察来发现世界的结构，而不是通过告知每个对象的名称来发现世界的结构。
人类的视觉是一个活跃的过程，它使用具有高分辨率，低分辨率的小中心凹，以智能的，针对特定任务的方式依次对光学阵列进行采样。我们预计，未来视觉的许多进步都将来自端到端经过培训的系统，并将ConvNets与RNN结合起来，后者使用强化学习来决定要看的地方。结合了深度学习和强化学习的系统尚处于起步阶段，但是在分类任务上它们已经超过了被动视觉系统，并且在学习玩许多不同的视频游戏方面产生了令人印象深刻的结果。

自然语言理解是深度学习必将在未来几年产生巨大影响的另一个领域。 我们期望使用RNN理解句子或整个文档的系统在学习一次选择性地关注一部分的策略时会变得更好。
最终，人工智能的重大进步将通过将表示学习与复杂推理相结合的系统来实现。尽管长期以来，深度学习和简单推理已被用于语音和手写识别，但仍需要新的范例来通过对大向量进行运算来代替基于规则的符号表达操作。