In the early and mid 2000, AI is something mythical that only tech giants could do, having access to big data and army of data scientists. However, in 2020, this is no longer the case. Many startups actually leverage this smart technology as a differentiation edge from the competition.
Many service providers offer a well-built intelligent service which you could leverage. These services provide a good and accurate result for many common use cases which you could use to enhance your features and user experience. In this article we will explore intelligent services that Azure is offering.
Azure offers 4 services around text analytics: sentiment analysis, key phrase extraction, language detection, and named entity recognition. You could try out the feature from https://azure.microsoft.com/en-au/services/cognitive-services/text-analytics/. These services are accessible via API calls, which mean that you could incorporate these to automate some of your processes or to enhance some existing features of your apps.
Now, let’s see how we could call these APIs:
- First, go to https://portal.azure.com/
- In the search box, type “text analytics”, and create it.
- And create a new service, fill in the details, and select a region where you want to deploy your service.
- After the deployment is complete, go to the resource.
- You will be provided with subscription key, which will be used to call the deployed service.
- Now you have a deployed text analytics service!
You could test the service you just deployed. Under Resource Management > Quick start > section 2, go to API console. Then, you want to select the region you deployed the service. You will be presented with a form that help you craft the API request. The most important field here is the Ocp-Apim-Subscription-Key, which refers to your service subscription key.
You could also call the service via other tools, such as curl or Postman. And of course, you can call it in your apps.
Let’s try another service that Azure provided: a service to moderate user submitted content, which includes text, image, and video content.
The steps to provision this service is similar to the previous steps:
- Create a new resource.
- Search for “content moderator” and create it.
- After it is successfully created, go to the resource.
- To try out the deployed service, go to “Content Moderator API Reference”.
The service provides a built-in feature to detect personally identifiable information (PII) and offensive content in text content. For image/video content, the service helps identify adult or racy content.
Azure provide a rich library of computer vision services. You could try some of the capabilities in: https://azure.microsoft.com/en-au/services/cognitive-services/computer-vision/.
Similar to the services discussed before, deploying Azure computer vision service is as simple as creating a new service in Azure, search for “computer vision”, and deploy the service. After the service is deployed, you could access the API console to try out the APIs.
Some interesting services offered:
- Analyze image: extracting various information from an image. Below is the information category that Azure cognitive service could extract.
- Describe image: using human natural language to describe an image.
- Get thumbnail: identifying area of interest in an image, and automatically crop the image.
- OCR: extracting text information from image.
- Object detection: tagging and describe objects found in an image.
AI is no longer limited to be something that only big tech companies, with huge investment, can do. There are many general purpose AI services which you could leverage, so you could focus your resources on developing core features of your business. As demonstrated in this article, it is very easy to incorporate intelligent capabilities to your apps or processes.