Advanced Cross-Translation Service

We are inspired by the amazing work of DeepL.com, the most sophisticated translation service out there. If necessary, we will leverage their API to assist building up the database.

We are starting out with a small scope and expand on it once we have the main set working nicely. Our focus is to enable accurate and authentic cross-translation for the ecosystem. In addition, we are including further detail on lexemes so that you can use the app either to translate or to get relevant information on a topic. This will help translators and every user in the community to engage with content.

We are leveraging the knowledge of the community to intuitively correct translations/choose the most appropriate translation out of different options. The community can flag unknown words/sentences and write them on a wish list, which assures we focus on the most needed content.

At this time, we are forming teams for fourteen different languages, which we think to be of special importance for the Cardano ecosystem. These are:

English, Mandarin, Arabic, Spanish, French, Japanese, Swahili, German, Amharic, Oromo, Indonesian, Russian, Hindi and Portuguese.

We do so for the game we are building right now: a social multiplayer game for language learning, sharing and collaboration. (link below)

Scope of this proposal

We want to create an intuitive and advanced cross-translation web-app for translating words and sentences. The first focus of this Dapp is content for basic communication and blockchain education.
To achieve this, we are starting with a fixed set of words + sentences, that are decided on in agreement with all fourteen language teams according to our intercultural translation concept. The language teams (assisted through DeepL) will provide the initial translation effort to build a common ground, and then slowly expand the database and connections in the graph network through community engagement.

The Sandbox-MMO for language learning will provide additional mid to long term growth of the database. It will add more content, translations, explanations/details, images and recordings.
https://cardano.ideascale.com/a/dtd/Sandbox-MMO-for-learning-languages/381663-48088

We also assist community efforts to connect and form digital hubs in different languages.
https://cardano.ideascale.com/a/dtd/Assisting-14-Digital-Language-Hubs/382940-48088

Who is behind this?

Educhainment is building tools and game-based learning applications on the Cardano blockchain. We got funded in Fund 6 to build a play-to-learn-and-earn game for learning languages. We are leveraging our existing linguistic efforts and research into intercultural translation + machine translation to provide resources for the community.

https://educhainment.com/

What are we asking funds for?

We want to create a dedicated team, that is only focusing on the creation and design of the translation Dapp. This includes the database and implementation of existing FOSS-approaches and training of neural networks for machine translation. The team will work in close collaboration with our Educhainment team. Educhainment is funded self-sufficiently, so we are only asking for funding for the new team.

We want to hire three more people for three months to kickstart and develop the backbone of the database. These will work part-time for ~60 hours on the project and will get $1200 per month, which equals $20/h. This is enough to at least cover the basic costs of living. This core team is responsible for:

1. Data optimization and machine learning
2. Training of neural networks with existing content
3. Language technology + translation/information science

Breakdown of Funds:

• Setting up servers for the database ($200)
• Team wage for three months (3 * 3 months * $1200 = $10800)
• Subscription to tools/DeepL-API ($500)

-> $11500

Roadmap:

• Now - Q1 2022: Formation of language teams. Machine learning + Linguistic research. Creation of intercultural translation concept. Gather core team for cross-translation app.
• Q2 2022: Train neural networks. Build database and create translation graph for lexemes. Design language-specific appropriate interfaces. Connect with existing translation and linguistic groups in the Cardano ecosystem.
• Q3 2022: Finished product. Continued community driven expansion of database and utilities

KPIs:

1. Number of words + sentences supported
2. Size of database
3. Number of languages supported
4. Monthly active users of translation service
5. Accuracy of blockchain content translation in %

IP:

We are leveraging open source solutions and are reliant on contribution of translation efforts by the community. So naturally, we are open source (MIT) and free to use!

What does success look like?

We created a simple and free to use cross-translation service, that is growing with the community needs. It is focused on the most relevant subjects of the Cardano ecosystem and their users to enable/enhance communication and get information.