Artificial Intelligence Platforms for developers: speech recognition

4 min reading
Digital transformation / 19 January 2016
Artificial Intelligence Platforms for developers: speech recognition
Artificial Intelligence Platforms for developers: speech recognition

BBVA API Market

Two words: Artificial intelligence (AI). It was one of the strongest fields of development during 2015 and it will also be in 2016. It will certainly be one of the sectors generating most profits for years or decades to come. Imagine an area of hundreds of hectares of arable land, but few planted seeds. Companies like Google, Facebook or Microsoft take time to develop Artificial Intelligence products, generally linked to natural language processing and speech recognition

Some estimates state that the market for Artificial Intelligence-related applications alone will be worth $11.1 billion by 2024. Others say that the global market for Artificial Intelligence will reach $40 billion by 2022, with an annual growth rate of approximately 25% and in sectors as diverse as transport or health. Whether they are right or not, the field of AI is a market for developing ideas, projects and companies with a huge projection.

Today there are several startups that have launched platforms and solutions for Artificial Intelligence so developers are able to create their projects. This is a list of some of the best options, especially in the field of apps, speech recognition and artificial intelligence: 

●      Jasper

Jasper is an open code platform for application developers with voice control. All kinds of products can be controlled with it: using your voice to request information on specific applications, managing your home, updating social networks, among other things. The code is available for developers who want to program their own modules.

This platform is specifically designed for hardware Raspberry Pi, specifically its Model B board. It also needs additional hardware such as a WiFi adapter that allows you to have an Internet connection and a USB microphone (the creators of Jasper tested the platform with the Akiro Kinobo USB microphone). It is also advisable to have a 4GB memory card and an Ethernet cable. Here is the full list of hardware necessary to develop with Jasper

In the platform’s documentation there are already some modules written by other developers for other applications showing possibilities. A Google Calendar module has been created for updating events. The code is available on Github. Another for Twitter which makes it possible to launch tweets, check what the most important trends are, direct messages, etc. Jasper also brings some default modules for Facebook notifications and Spotify controls. Prerequisite: know Python

●      Api.ai

Api.ai is one of the most promising startups in natural language processing and speech recognition. It provides developers with all kinds of Siri-style conversational interfaces to incorporate a smart wizard into the applications. In mid-2015 it achieved a round of additional funding of $3 million of funds such as SAIC Capital. Previously companies like Intel or Motorola opted for them. 

This round of funding was due to the growth of its business purpose. Besides having interfaces to connect devices, it is now also committed to smart wizards that allow cars to be connected with platforms or services. It is the smart car that is already a reality.

Api.ai has Software Development Kits (SDK) for various platforms: operating systems such as Android, iOS or Windows Phone; Apple Watch; JavaScript interpreter on the Node.js server side; the framework of developing Cordova applications; 3D Unity videogame development environment; or programming languages like C#, Python or JavaScript. It can also be connected with other assistants such as Amazon Echo or Microsoft Cortana.

Api.ai has a completely free version, but also offers a fee-based version for companies: it provides private space in the cloud, supported by a professional team 24 hours a day, seven days a week, and speech recognition models that are customizable by the customer.

●      Wit.ai

They are defined as “natural language processing for developers” engaged in designing applications, robotics, wearables (connected wearable devices) or automatic messaging platforms. It is a really interesting startup. So much so that Facebook noticed it and bought it only a year and a half after its launch. At that time, 6,000 developers relied on their APIs to develop services; today that number has now reached 10,000.  Wit.ai is an open and free source code platform for professionals.

Wit.ai is able to process natural language and transform all that to the JSON data exchange format (JavaScript Object Notation), which is what developers ultimately work with for their applications. Wit.ai not only facilitates language recognition but separating each concept into entities and values, as well as a value of confidence in their recognition.

// get a response
{
  “_text”: “Turn on the lights”,
  “intent”: “lights”,
  “entities”: {
    “on_off”: “on”
  },
  “confidence”: 0.83
}

Wit.ai has an API for developers available for numerous operating systems, hardware and programming languages: Android, iOS, Windows Phone, Node.js, Raspberry Pi, Ruby, Python or C. It also has a JavaScript plugin for the development of front-end for application design.

●      Zypr

Zypr is a set of Web APIs released by Pioneer for the work of developers with speech recognition, but also with maps, routes, or social networks. Some of the services provided by Zypr include Facebook and Twitter Control; interaction with Google services like Gmail (mail service), Calendar (calendar) or Contacts (contacts); or controlling a restaurant recommendation application like Yelp.

Zypr APIs work with iOS and Android operating systems.

Follow us on @BBVAAPIMarket

It may interest you