SELMA (stream learning for multilingual knowledge transfer) and ESPERANTO (speech research and technologies) both focus on NLP, multilingualism, and neural network-based speech processing.
AVIGNON UNIVERSITE
French university specializing in multilingual NLP, speech technologies, cultural heritage science, and Mediterranean bioactive compound research.
Their core work
Avignon Université is a French university with strong capabilities in natural language processing, speech technologies, and multilingual AI — anchored by its computer science research labs. Alongside this digital expertise, the university contributes to cultural heritage science, archaeological materials research, and bioactive compound analysis from Mediterranean crops like olives. Their H2020 portfolio reflects a mid-sized institution that brings specialized NLP and speech processing skills to large European research networks, while maintaining roots in regional heritage and agri-food science.
What they specialise in
ED-ARCHMAT (European Doctorate in archaeological materials science) and Geopark (heritage, education, sustainable development) demonstrate sustained engagement in heritage preservation and digital techniques for archaeology.
Olive-Net investigated bioactive compounds from Olea europaea for food, cosmetic, and pharmaceutical applications including extraction, profiling, and nutraceuticals.
UCOM (Ultrasound Cavitation in Soft Materials) covers CFD simulation, bubble dynamics, and ultrasound-based cleaning — a physics-oriented niche for the university.
How they've shifted over time
In their earlier H2020 period (2015–2018), Avignon Université focused on Mediterranean agri-food science (olive bioactives, nutraceuticals, cosmetics) and cultural heritage education through geoparks and archaeological training. From 2021 onward, the university pivoted sharply toward AI-driven language technologies — multilingual NLP, stream learning, speech-to-text, and neural network explainability. This shift signals a deliberate investment in computational linguistics and AI as the university's flagship research direction for European collaboration.
Avignon is consolidating around multilingual AI and speech processing, making them a strong candidate for future projects needing low-resource language technologies or human-in-the-loop NLP systems.
How they like to work
Avignon Université exclusively participates as a partner — they have never coordinated an H2020 project, which is typical for a smaller French university contributing specialist expertise to larger consortia. With 84 unique partners across 28 countries from just 6 projects, they work in broad, diverse networks rather than tight recurring clusters. This makes them an accessible, low-friction partner comfortable integrating into new consortia without needing a leadership role.
Despite only 6 projects, Avignon has built a remarkably wide network of 84 partners across 28 countries, driven largely by participation in MSCA mobility actions that span many institutions. Their reach is genuinely pan-European with no heavy geographic bias.
What sets them apart
Avignon's distinctive value lies in combining multilingual NLP and speech processing expertise with deep roots in Mediterranean cultural heritage and agri-food science — a rare interdisciplinary mix. For consortium builders, this means a single partner that can contribute both AI/language technology work packages and domain knowledge in heritage digitization or food science. Their MSCA track record (5 of 6 projects) also signals strong doctoral training capacity and researcher mobility infrastructure.
Highlights from their portfolio
- SELMALargest funding (EUR 599K) and most technically ambitious project — stream learning for multilingual NLP covering summarization, machine translation, and speech-to-text.
- ED-ARCHMATA European Joint Doctorate (EUR 526K) in archaeological materials science, reflecting significant investment in training the next generation of heritage scientists.
- ESPERANTOFocused on speech processing explainability and low-resource languages — addresses a critical gap in AI transparency for underserved language communities.