The Project

  • Highest Security Standards

  • Conversational Artificial Intelligence

  • Natural Language Processing

  • Breaking language barriers

The goal of the project is to build a prototype for a personal virtual assistant, which can be installed on a small hardware device or integrated within the ELG container. We want to demonstrate that this can be accomplished using ELG language resources and technologies while keeping the highest security standards. We use open source software E.D.D.I which is running in a docker container for the natural language interface.

Main project goal is the creation of an ecosystem for conversational AI that leverages the ELG cloud infrastructure to provide IT project managers, business analysts, software developers and software architects with a simple mechanism to integrate conversational-based software solutions independently from the end user spoken language and/or the end point interpretation language.

We believe that conversational AI apps are the best to push the interactions between people that speak different languages due to their real-time nature and the ability to create personalized customer experiences at scale.

Use Cases

The use cases of YouTwindi personal virtual assistance
are countless, yet, they all serve the same purpose of

  • represent the user's needs...

    ... to the outside world by understanding the user's preferences, it's sentiment and thus act adequately on behave of the user

  • Ensuring privacy...

    ... by saving all the personal data on the user’s mobile device, secured by Verified Identification (biometric authentication) accessible via Text and Speech Recognition

  • Allow fast communication...

    ... with thousands of communication partners with bot 2 bot communication

Prototype I

Local News in Foreign Languages & fake news alert

We know that immigrating to a foreign country is often a challenging and life-altering experience. And learning a new language can be an even more tiring and time-consuming process! Our Prototyp will be the new place to go when it comes to learning a new language through the news of your new country!

We make trending local news, written in the national language, accessible to speakers of other languages, by presenting them with adjacent previews of both the original and the translated version.

This offers people an amazing opportunity to learn the language of their new home, effectively  using real language texts, rapidly enriching their vocabulary with new words and phrases. At the same time, users can contribute to the constant improvement of our Prototyp by correcting and submitting possible errors in translation. By doing so, this helps upgrade the translation algorithms to improve future translation results.

YouTwindi will remember the news you had consumed before and will watch the web for fake news alerts for that piece of information you previously consumed and will inform you about it.

Prototype II

What to gift a friend for birthday?

A perhaps quite simplistically looking use case, but yet technologically challenging is the task of:

“Finding a proper birthday gift for someone,
without knowing what they want and no person to ask for”

The solution for this problem can be personal assistants negotiating for a piece of information. Your bot could ask the bot of your friend what to gift the friend based on that amazon wish list (which is provided by Alexa).

As your and your friend’s bot are “friends” by themselves (trusted domain), they are allowed to communicate such information without your friend getting notified by it.

Technology

  • Docker containers running on Android

  • Peer-to-Peer Network for bot communication

  • Public/Private section of the accessible data

  • Identity check mechanisms to verify both users of the bot

  • Key Exchange and End-To-End Encryption

We are using a docker based open source bot engine E.D.D.I, built in java and uses mongodb as storage. In order to keep the data private, the bot must run fully on the mobile device.

As mongodb and E.D.D.I are running on docker, we are running our project also as docker containers on Android.

For the virtual assistant prototype we have decided to run a virtual linux via QEMU, an Alpine Linux on an Android device. This enables us to have the mongodb and the E.D.D.I fully running on the device.

The main focus of the project is to create a prototype that does bot-to-bot communication. To implement that, we will create a peer-to-peer network structure, where every peer is a YouTwindi instance that runs as a docker container on a mobile device. These instances have a public part of the bot. The public bot gives out all information that the user deems to be public. It can be customized, but at least there will be a handshake, where two YouTwindi instances can negotiate whether their users actually want to share private information. The handshake process is an identity check on both sides. If both identities have been successfully verified, the bots exchange keys and an end-to-end encrypted bot-to-bot communication commences.

Work Plan

  • Research

    Software Research | Hardware Research | Use Cases definition

  • Software Integration

    Integration EDDI and ELG | Container Setup | Use Case Implementation | Testing

  • Hardware Integration

    Preparation of Hardware | Installation of Software | Integration Tests

  • Final protoyping

    Finalization of Prototype | Final Tests | Prototype presentation preperation

  • Dissemination

    Set-up project Website | Ongoing information updates via Website | Social Media Posts

100%

80%

100%

75%

90%

with 24 official EU and many more additional languages, multilingualism in Europe and an inclusive Digital Single Market can only be enabled through Language Technologies

In Cooperation

E.D.D.I by Labs.ai

A, over 12 years developed, chatbot framework, which is open source and can be used to create, maintain and run multiple chatbots. Our experience is related to natural language processing, which is at the core of artificial intelligence and conversation design.

European Language Grid

The European Language Grid will foster Language Technologies for Europe built in Europe, tailored to our languages and cultures and to our societal and economical demands, benefitting the European citizen, society, innovation and industry.

Send us a message