认知服务和机器学习Cognitive Services and machine learning

认知服务提供机器学习功能以解决常见问题，例如分析文本以获取其情感情绪，或分析图像以识别物体或人脸。Cognitive Services provides machine learning capabilities to solve general problems such as analyzing text for emotional sentiment or analyzing images to recognize objects or faces. 无需机器学习或数据科学方面的专业知识就能使用这些服务。You don't need special machine learning or data science knowledge to use these services.

认知服务是一组服务，每个服务都支持不同的通用预测功能。Cognitive Services is a group of services, each supporting different, generalized prediction capabilities. 服务分为不同类别，以帮助你找到合适的服务。The services are divided into different categories to help you find the right service.

服务类别Service category 目标Purpose
决策Decision 构建应用，用于呈现有助于做出明智和高效决策的建议。Build apps that surface recommendations for informed and efficient decision-making.
语言Language 允许应用使用预建脚本处理自然语言、评估情绪及学习如何识别用户想要的内容。Allow your apps to process natural language with pre-built scripts, evaluate sentiment and learn how to recognize what users want.
语音Speech 将语音转换为文本，将文本转换为自然语音。Convert speech into text and text into natural-sounding speech. 从一种语言翻译成另一种语言，并启用说话人验证和识别。Translate from one language to another and enable speaker verification and recognition.
影像Vision 识别和确定你的图片、视频和数字墨迹内容，为它们添加描述文字和编制索引，并审查这些内容。Recognize, identify, caption, index, and moderate your pictures, videos, and digital ink content.

在以下情况时使用认知服务：Use Cognitive Services when you:

  • 可以使用通用解决方案。Can use a generalized solution.
  • 从编程 REST API 或 SDK 访问解决方案。Access solution from a programming REST API or SDK.

在以下情况时使用另一种和机器学习解决方案：Use another machine-learning solution when you:

  • 需要选择算法并需要针对非常具体的数据进行训练。Need to choose the algorithm and need to train on very specific data.

什么是机器学习？What is machine learning?

机器学习是一种概念，你可以将数据与算法结合在一起，以解决特定需求。Machine learning is a concept where you bring together data and an algorithm to solve a specific need. 对数据和算法进行训练后，输出是可以再次用于其他数据的模型。Once the data and algorithm are trained, the output is a model that you can use again with different data. 已训练的模型可根据新数据提供见解。The trained model provides insights based on the new data.

构建机器学习系统的过程需要一些机器学习或数据科学方面的知识。The process of building a machine learning system requires some knowledge of machine learning or data science.

机器学习是使用 Azure 机器学习 (AML) 产品和服务提供的。Machine learning is provided using Azure Machine Learning (AML) products and services.

什么是认知服务？What is a Cognitive Service?

认知服务提供了机器学习解决方案中的部分或全部组件：数据、算法和训练模型。A Cognitive Service provides part or all of the components in a machine learning solution: data, algorithm, and trained model. 这些服务旨在要求你具备有关数据的一般知识，而无需具备机器学习或数据科学方面的经验。These services are meant to require general knowledge about your data without needing experience with machine learning or data science. 这些服务同时提供 REST API 和基于语言的 SDK。These services provide both REST API(s) and language-based SDKs. 因此，你需要具备编程语言知识才能使用该服务。As a result, you need to have programming language knowledge to use the services.

认知服务和 Azure 机器学习 (AML) 的相似之处是什么？How are Cognitive Services and Azure Machine Learning (AML) similar?

两者的最终目标都是应用人工智能 (AI) 以增强业务运营，不过，它们在各自的产品/服务中提供此服务的方式有所不同。Both have the end-goal of applying artificial intelligence (AI) to enhance business operations, though how each provides this in the respective offerings is different.

通常，受众不同：Generally, the audiences are different:

  • 认知服务适用于没有机器学习经验的开发人员。Cognitive Services are for developers without machine-learning experience.
  • Azure 机器学习是为数据科学家量身定制的。Azure Machine Learning is tailored for data scientists.

认知服务与机器学习有何不同？How is a Cognitive Service different from machine learning?

认知服务提供已训练的模型。A Cognitive Service provides a trained model for you. 该模型将数据和算法结合在一起，可从 REST API 或 SDK 获得。This brings data and an algorithm together, available from a REST API(s) or SDK. 你可以在几分钟内实现此服务，具体取决于你的方案。You can implement this service within minutes, depending on your scenario. 认知服务提供针对常见问题的答案，例如文本中的关键短语或图像中的项目标识。A Cognitive Service provides answers to general problems such as key phrases in text or item identification in images.

机器学习是一个通常需要较长时间才能成功实现的过程。Machine learning is a process that generally requires a longer period of time to implement successfully. 这些时间花费在数据收集、清理、转换、算法选择、模型训练和部署上，以达到认知服务所提供的相同级别的功能。This time is spent on data collection, cleaning, transformation, algorithm selection, model training, and deployment to get to the same level of functionality provided by a Cognitive Service. 通过机器学习，可以为高度专业化和/或特定的问题提供答案。With machine learning, it is possible to provide answers to highly specialized and/or specific problems. 机器学习问题需要熟悉相关问题的特定主题和数据，以及数据科学方面的专业知识。Machine learning problems require familiarity with the specific subject matter and data of the problem under consideration, as well as expertise in data science.

你拥有哪种类型的数据？What kind of data do you have?

认知服务作为一组服务，对于已训练的模型可能不需要、需要部分或需要全部自定义数据。Cognitive Services, as a group of services, can require none, some, or all custom data for the trained model.

无需其他训练数据No additional training data required

提供完全训练的模型的服务可被视为不透明框。Services that provide a fully-trained model can be treated as a opaque box. 你无需知道它们的工作方式或用于训练它们的数据。You don't need to know how they work or what data was used to train them. 你可以将数据引入已完全训练的模型，以获得预测。You bring your data to a fully trained model to get a prediction.

需要部分或全部训练数据Some or all training data required

某些服务允许引入自己的数据，然后训练模型。Some services allow you to bring your own data, then train a model. 这样，就可以使用服务的数据和算法通过自己的数据来扩展模型。This allows you to extend the model using the Service's data and algorithm with your own data. 输出可以满足你的需求。The output matches your needs. 导入自己的数据时，可能需要以特定于服务的方式标记数据。When you bring your own data, you may need to tag the data in a way specific to the service. 例如，如果要训练模型以识别花朵，则可以提供花朵图像的目录以及每个图像中花朵的位置以训练模型。For example, if you are training a model to identify flowers, you can provide a catalog of flower images along with the location of the flower in each image to train the model.

服务可能允许提供数据以增强其自身的数据。A service may allow you to provide data to enhance its own data. 服务可能需要提供数据。A service may require you to provide data.

需要实时或准实时的数据Real-time or near real-time data required

服务可能需要实时或准实时数据来构建有效的模型。A service may need real-time or near-real time data to build an effective model. 这些服务处理大量的模型数据。These services process significant amounts of model data.

数据模型的服务要求Service requirements for the data model

以下数据根据服务允许或需要的数据类型对每种服务进行了分类。The following data categorizes each service by which kind of data it allows or requires.

认知服务Cognitive Service 无需训练数据No training data required 你提供部分或全部训练数据You provide some or all training data 实时或准实时数据收集Real-time or near real-time data collection
计算机视觉Computer Vision xx
内容审查器Content Moderator xx xx
人脸Face xx xx
语言理解 (LUIS)Language Understanding (LUIS) xx
语音文本转语音 (TTS)Speech Text-to-speech (TTS) xx xx
语音语音转文本 (STT)Speech Speech-to-text (STT) xx xx
语音翻译Speech Translation xx
文本分析Text Analytics xx
翻译Translator xx

*个性化体验创建服务仅需要服务（在实时运行中）收集的训练数据来评估策略和数据。*Personalizer only needs training data collected by the service (as it operates in real-time) to evaluate your policy and data. 个性化体验创建服务不需要大型历史数据集即可进行前期或批量训练。Personalizer does not need large historical datasets for up-front or batch training.

在哪里可以使用认知服务？Where can you use Cognitive Services?

这些服务可在任何可进行 REST API 或 SDK 调用的应用程序中使用。The services are used in any application that can make REST API(s) or SDK calls. 应用程序的示例包括网站、机器人、虚拟或混合现实、桌面和移动应用程序。Examples of applications include web sites, bots, virtual or mixed reality, desktop and mobile applications.

如何使用认知服务？How can you use Cognitive Services?

每个服务都提供有关你的数据的信息。Each service provides information about your data. 可以将服务组合在一起以形成链式解决方案，例如将语音（音频）转换为文本，将文本翻译为多种语言，然后使用翻译后的语言从知识库中获取答案。You can combine services together to chain solutions such as converting speech (audio) to text, translating the text into many languages, then using the translated languages to get answers from a knowledge base. 可以直接使用认知服务创建智能解决方案，也可以将其与传统的机器学习项目相结合，以补充模型或加速开发过程。While Cognitive Services can be used to create intelligent solutions on their own, they can also be combined with traditional machine learning projects to supplement models or accelerate the development process.

了解详细信息Learn more

后续步骤Next steps