With MOVEN, Redlink provides a technology to easily distribute and version training models for machine learning on a proven infrastructure. Such models are used in AI powered solutions like chatbots, recommendation engines and recognition for named entities and sentiments from textual content.

About the Innovator

Redlink GmbH was founded in March 2013 and is headquartered in Salzburg, Austria. Redlink leverages the power of Apache Stanbol, Apache Marmotta and Apache Solr to deliver enterprise-grade content enrichment and linked data solutions. With our platform, we deliver semantic technology as service for national and international clients. As stand-alone solutions we power intranet searches, chatbots and content-based recommendations as well as data integrations and their presentation on the web or as mobile apps.

/futurium/en/file/redlinkjpgredlink.jpg

What is the innovation

In the context of the EU-funded SSIX project MOVEN (“models” + “maven”) was conceived to to provide the (up to then, non-existing) means for easily distributing and versioning large training models for Machine/Deep Learning. The result is a model-agnostic tool, making advantage of the features of one of the largest Apache projects, Maven. Moven is currently in a prototype stage but fully functional and compatible with Java and Python projects. Its development is driven mainly by SSIX project needs.

Out of the lab - Into the Market

Several of Redlink’s products and services use models for supporting NLP tasks, e.g. our platform, the search, recommender solutions as well as our backends for chatbots. In agile settings like chatbots, where training of models is an ongoing, highly automated task, Moven provides this facility for us. In general, it let us compete with big players in the market of AI and machine learning.

Benefits of participation in the Framework Programme

H2020 helped Redlink to gain insights in excellent research with academic partners in sentiment analysis for the financial domain. We could provide our expertise to set-up a processing pipeline for social media messages to extract named entities and getting trends for sentiments on companies and products. Its research context and agenda directly fuels our product development roadmap - using technology and research insights for data integration and semantic extraction.