Machine learning is about teaching machines to recognize patterns like humans do. While even a two-year-old can easily distinguish between an apple and an orange, it is very difficult to teach a computer to do the same thing. Google has developed pre-trained machine learning models and made them available to you as easy-to-use Google Cloud APIs. Now, with just a few lines of code, you can add artificial intelligence or AI to your own application. In this module, we'll explore Google's pre-trained machine learning APIs for vision, speech, video intelligence, and natural language processing. Google Cloud Platform offers several pre-trained machine learning models that you can use to add intelligence to your application. Cloud Vision API enables you to perform complex image detection. Cloud Speech API enables developers to convert audio to text. It handles 110 languages and variants to support your global user base. You can transcribe the text of user's dictating to an application's microphone, enable command and control through voice, transcribe audio files and more. Cloud Video Intelligence API enables you to search every moment of your video file, to extract and understand the video's entities at the shot, frame, or video level. The API annotates videos stored in Cloud storage and helps you identify key noun entities of your video and when they occur within the video. Cloud Translation API enables you to translate an arbitrary string into any supported language. Translation API is highly responsive. Websites and applications can use Translation API for fast, dynamic translation of text from a source language to a target language. Cloud Natural Language API enables you to extract information about entities such as people, places, and events that are mentioned in text documents, news articles or blog posts. You can use the API to understand sentiment about your product on social media or pass intent from customer conversations. You can also use your own data to build and train your own machine learning models by using TensorFlow and Cloud AI platform. It's really easy to invoke the REST APIs to implement machine learning in your application, no ML knowledge is required. Let's take a look at this example. In this example, we are using the Vision API to process an image that's stored in Google Cloud Storage. We invoke the REST API and send it a JSON request, and we receive a JSON response with attributes that describe the image, as simple as that. Let's take a look at a few examples now. The Vision API can categorize objects under labels and perform optical character recognition or OCR. The Vision API can detect landmarks, logos, faces, and explicit content. For example, Vision API can analyze faces and return information about the emotions and headwear. In the wedding picture, the API accurately returns the emotional expressions on the faces in the picture and it says that it's highly likely that the expression is one of joy. In the picture of the sphinx, the Vision API correctly detects that the image is from the sphinx in Las Vegas and not the sphinx in Egypt, pretty cool. With the Speech API, you can transcribe the text of users dictating to an application's microphone, enable command and control through voice, transcribe audio files and more. At Google, we have a problem where people might book a conference room for a meeting, but then not make it in to the room in person. What happens then is that the conference room is booked and cannot be used by somebody else who really needs the space. At Google, we decided to solve the problem with our own tools. Google's conference room systems perform occupancy detection by using motion detection with the video conference camera and by caller ID matching. Every 30 seconds, the video camera unit sends a Cloud Pub/Sub notification indicating whether motion was detected or not. It also sends a Cloud Pub/Sub notification when a call starts or ends. If, say, motion is detected between six and eight minutes after the meeting start time, the room counts as occupied. Otherwise, it's empty and can be released to others for use. Imagine the possibilities with the pre-trained machine learning APIs. How can you add intelligence to your application?
