Chatbots expansion within the last couple months is continuing to change the direction of the tech world development towards the most popular messaging platforms, like Facebook Messenger, Telegram, WeChat, Line, Skype, Twitter Direct Messages, Slack, Google Assistant and others. However, chatbot experience might not be that exciting if the chatbot you are talking to is too basic and has nothing to do with advanced AI or natural language processing tools.
In this article I would like to focus mainly on major AI tools that are currently available on the wb and how you can use them when developing a smart chatbot.
As we all know Google Assistant is one of the smartest chatbots available for millions of Android and iOS users. And no wonder, as being a tech giant, Google can make use of its powerful AI engine that empowers Google search abilities and is the language understanding system behind the Google Assistant.
Let’s take a deeper look at Cloud Natural Language API that derives insights from unstructured text using Google machine learning. Its major capabilities include the following:
1) Syntax Analysis
Syntax analysis offered by Google provides a possibility to identify the structure and meaning of the text with the help of machine learning models in REST API, identify parts of speech, extract necessary information from the sentences and create dependency threads for each sentence. It can be used to extract information about people, different places and events in the text documents, articles or blogs. Syntax analysis also allows understanding sentiment about a certain product, feature and parse intent from the chatbot users.
2) Entity Recognition
Entity Recognition makes it possible to identify entities by their types such as person, organization, location, events, media, etc. and receive insights on product reception and user-chatbot overall experience.
3) Sentiment Analysis
This feature makes it possible to understand the sentiment expressed in a block of text: positive, negative or neutral.
Multi-Language helps you to analyze texts in multiple languages such as: English, Spanish and Japanese using Translate API. Your app API can also be combined with Google Cloud Speech API to extract insights from audio conversations and Vision API OCR to understand scanned documents.
5) Integrated REST API
Cloud Natural Language API provides access via REST API. Any text file can be uploaded in the request or integrated with Google Cloud Storage.
Google Cloud Platform is free for projects that require up to 5K units processing monthly.
Another smart solution to empower your bots with intelligence is offered by Microsoft Bot Framework. It provides a possibility to build cross-platform chatbots using the SDKs for Node.js, .NET or REST. All bots can be channeled to such platforms as Slack, Telegram, GroupMe, Skype, Messenger, Kik, Twilio. The platform also offers to Visit the Bot Directory to try out available bots, and add them to your favorite conversation experiences.
One of the most exciting features of the platform is Cognitive Services which can be used to make the bot smarter and interpret user inputs in a more meaningful way. Cognitive services will make the bot more intelligent which makes a great difference. The bot will be able to understand almost everything from the natural language and respond with phrases that make sense. To provide useful answers the bot will extract knowledge from different sources and at the same time learn as it acquires more experiences which will be used to improve its capabilities.
Microsoft Cognitive Services are based on a huge collection of powerful AI algorithms which can be added into the bot with just a few lines of code. The APIs can be integrated into a bunch of languages on the platform you need or choose. Microsoft provides comprehensive documentation and API references on the Cognitive Services documentation site. You will also find code samples there to start developing bots with cognitive skills with flying colours.
LUIS (Language Understanding Intelligent Service) is part of the Microsoft Cognitive Services package. It is in beta now and is free to use. It supports such browsers as Internet Explorer 10/11 and Chrome.
This platform offers a wide variety of possibilities such as:
- create models for your chatbot to better understand user inputs or certain entities;
- use pre-built, world-class models from Bing and Cortana to recognize entities like places, times, etc.;
- deploy models to an HTTP endpoint with one click and receive easy-to-use JSON in return;
- activate language understanding models from a chatbot on any device.
Microsoft Cognitive Services package is available for a free trial plan. You just need to have a Microsoft Account set up and have access to your subscriptions page to view the products you are using, your quota left and the ability to add additional products to your subscription.
As we can see the world of smart bots is just a question of the nearest years. As AI and NLP are developing and progressing day-by-day, this is not a phantasy any more but an evolving future.