Machine learning is one of the current hot topics, with more and more companies and products embracing it. AWS provides a great platform for machine learning and has some amazing products to help users with various areas, such as speech recognition, video recognition, translation, transcription, and others. Let's look at some of AWS's machine-learning products in more detail:

• Amazon SageMaker: This is a platform for machine learning. This service enables data scientists to build their machine-learning models, train the models with sample data, and then deploy the model to production. The models are refined later, based on the inferences generated and the new data obtained in production. SageMaker has hosted Jupyter notebooks, integrated with S3, which allows developers to visualize their data that's stored in S3. SageMaker also provides the most common ML algorithms. It can run TensorFlow and Apache MXNet in Docker containers. When users want to train their models, SageMaker makes it possible by easily scaling up to petabyte scale. Once trained, the model can be deployed using SageMaker on an auto-scaling cluster. This process is shown in the following diagram:

• Amazon Lex: This is a service that enables developers to build chatbots using both voice and text interfaces. Lex utilizes deep-learning functionalities of automatic speech recognition to convert speech to text. Lex uses natural-language understanding to understand the intent of text. All these factors ensure that the chatbots are lifelike. With Lex, developers are now able to access the same deep-learning technologies that were used to build Amazon Alexa.
• Amazon Polly: This is part of the language services provided by AWS using machine learning. AWS Polly is a text-to-speech conversion service. Deep learning is used to produce speech that is very close to a human voice.
• Amazon Translate: This is another language service from Amazon. This service is a neural machine translation service that translates supported languages and English. Using deep-learning technologies, it provides high-quality translations. Currently, six languages are supported: Arabic, Simplified Chinese, French, German, Spanish, and Portuguese. More languages will be added in due course.
• Amazon Transcribe: This is a speech-to-text service. This service is very useful in cases where voice calls need to be transcribed or when subtitles need to be added to videos. Amazon Transcribe uses the audio files stored in S3, analyzes them, converts the speech to text, and returns a text file to the user, containing the transcribed speech.
• Amazon Comprehend: This is a natural-language processing service used to discover relationships and insights in text. For example, Amazon Comprehend can scan a lot of text—say, from multiple websites—and find out how a particular product has been rated. Using machine learning, Amazon Comprehend can identify key words, understand how positive or negative the text is, identify brand names or people's names, and much more. The results of Amazon Comprehend can be used in a variety of applications, be it analyzing customer support incidents, identifying which feature of your product your customers love the most by analyzing social media feeds, and so on. 
• Amazon Rekognition: This is a video and image-analysis service. This service enables the recognition of people, text, activities, objects, and so on. Rekognition can be used in various tasks and services, such as comparing faces for facial recognition, user verification, people counting, and so on. It can also be used to detect inappropriate content. Rekognition uses image files and video files stored in S3 and analyzes them.