AI WEEKLY SECTION: DISCOVER OPEN SOURCE PROJECTS FOR ARTIFICIAL INTELLIGENCE

We have already mentioned in previous posts several benefits on how Artificial Intelligence (AI) can favour society: companies, health and the environment.

In this post, from the core of OpenExpo Europe, we show you some of the AI ​​platforms that are still on Open Source.

1. Acumos – The Linux Foundation

Acumos is part of the LF Deep Learning Foundation, an organisation part of The Linux Foundation, which is responsible for maintaining innovation of the AI ​​within Open Source, making these technologies available to developers, scientists and data researchers.

This platform allows AI applications to be created, exchanged and established, standardising the infrastructure stack and the tools needed to run an AI environment ready to use. It aims in freeing developers and data scientists, so they can focus on their core competencies and accelerating innovation.

Among the advantages that it includes there are: tool packages such as TensorFlow and SciKit Learn and models with a common API, allowing an easy connection; the admission of popular software languages ​​such as Java, Phyton and R; and a catalog of AI models provided by the community that can be shared safely.

You can find the link here to the Acumos code.

2. Microsoft Cognitive Toolkit (CNTK).

The “technological giant” could not be absent, asi its AI platform. In this case, it is an AI solution that allows taking Deep Learning projects to the next level.

You can find here the link to the code, which is available on GitHub.

Find a video published by Microsoft so that you can understand its operation:

 

3. Jasper

It is an open source platform for application developers with voice control.

The application allows you to use the voice to request any information, update social networks and control home appliances, among others. Likewise, Jasper is always listening, being able to pick up the user’s voice a few meters away.

Jasper is designed specifically for a specific hardware, the Raspberry Pi Model B. In addition, you need additional hardware such as a WiFi adapter, a USB microphone (here is a link to the adapter used by the creators of Jasper), a 4GB memory card and a Ethernet cable. Here is the full list.

Have a look at the available code on GitHub. It requires knowledge of Python and HTML.

Below you can find an introductory video of its creators.

4. Wit.ai

Wit.ai, acquired by Facebook, is an open source platform that is defined as “natural language for developers” and according to its website is already used by more than 180,000 professionals.

This application makes it easier for developers to create tools and devices with which users can talk or send messages. It is available for bots, mobile applications, wearables and for homes.

The platform is able to process the natural language and transform it into JSON (JavaScript Object Notation), the language that developers use for applications. Also, it has an API for developers available for: Android, iOS, Windows Phone, Node.js, Raspberry Pi, Ruby, Python or C.

Here is the link to the code in Open Source, available on GitHub.

So that you do not miss out on other similar platforms, have a look at their links: Caffe, TensorFlow, Torch, Scikit-Learn, Keras and Theano.

Are you interested in more news and trends on AI?

If you are, you cannot miss out the AI ​​& IOT Summit, a co-located conference on June 20th.

Success stories will be presented, as well as their application in different sectors, such as Machine Learning, Legal Tech, Language Processing, Robotics, Augmented Reality, Cognitive Computing, among others. The presentations will be done by Keynote Speakers, experts and professionals.