Rome’s Translated leads new €29M Horizon Europe project that aims to bring AI into the real world

Rome-based Translated, a firm specialised in AI-powered language solutions, has officially announced it will lead DVPS (Diversibus Viis Plurima Solvo, Latin for “Through diverse paths, I solve many issues”), a €29M Horizon Europe initiative. 

DVPS aims to create AI systems that learn from text, images, and real-world experiences, combining language, vision, and sensor data for a better understanding of the world.

The initiative brings together 20 leading organisations across 9 countries, with a core team of 70 top-tier European AI researchers.

Marco Trombetti, CEO of Translated, comments, “Large language models have been a breakthrough, but we’re starting to see their limits both in their architecture and how they learn from static, human-made content, only available in the digital world. To go further, AI must interact with the real world in real-time. With DVPS, we’re enabling machines to grow by engaging with reality, and exchanging their learning between themselves instantly.”

The London-based Alan Turing Institute is taking the lead role in the Geo-Intelligence workstream of the DVPS project. 

With this move, the Turing Institute will explore how multi-modal foundation models can enhance Earth system forecasting.

Solving real-world challenges with multimodal intelligence

Today’s AI translation systems struggle in crowded, noisy environments. DVPS tackles this limitation by fusing visual, spatial, and acoustic data to determine which speaker to focus on, delivering far more accurate translations, even in chaotic conditions.

For example, they often fail to provide live translation in busy and noisy places with multiple speakers.

Traditional systems have difficulty identifying who is speaking, while humans can easily focus on the right person by using visual and sound cues, such as where someone is looking, their facial orientation, and how loudly they are speaking.

This technology has various potential applications, such as:

• Language: Allowing real-time translation in many languages while understanding text, speech, gestures, and the surrounding context.
• Health: Helping detect cardiovascular risks early by creating a 3D digital model of the heart from medical images.
• Environment: Improving disaster response through flood predictions using satellite and drone data, along with real-time observations.

A European collaboration at scale

The initiative is led by Translated, which oversees the overall vision and execution.

The DVPS founding team consists of 70 top European AI scientists from across the following partners:

• Research partners include the University of Oxford, ETH Zurich, Imperial College London, The Alan Turing Institute, EPFL, Universitat de Barcelona, Karlsruhe Institute of Technology, and more.
• Vertical partners span Heidelberg University Hospital, Vall d’Hebron Research Institute, Amsterdam UMC, and leading AI companies like Deepset, Sistema, and MEEO.
• HPC support comes from Cyfronet, Poland’s national supercomputing center.

This initiative not only strengthens Europe’s leadership in AI but also helps bridge the gap between digital intelligence and human-like perception.

Translated: Providing AI-powered language solutions

Translated is a language services provider that uses artificial intelligence (AI) to help professional translators.

Founded in 1999 by linguist Isabelle Andrieu and computer scientist Marco Trombetti, the company aims to ensure everyone can communicate in their own language.

In 2012, they introduced context-adaptive machine translation.

Trusted by major translation clients, Translated has received recognition for its AI technology from organisations like IDC Marketscape, CSA Research, and Gartner.

In 2024, they launched Lara, an innovative translation AI that rivals the quality of skilled human translators. By combining insights from over 500,000 native translators with machine efficiency, Lara makes complex localisation projects possible.

Currently, Translated provides fast and high-quality translations to over 300,000 customers in 200 languages and 40 subject areas.