CENTRO DE APLICACIONES TECNOLOGICAS DE AVANZADA

Their core work

CENATAV is a Cuban applied AI research center based in Havana, specializing in computer vision, speech processing, and natural language processing. Their work has covered two distinct but related domains: biometric and forensic image analysis (person identification from multimedia) and spoken/written language technologies, with a particular emphasis on low-resource language settings. As a third-country institution participating in EU research through MSCA-RISE staff exchanges, they contribute specialized AI expertise — especially relevant for Spanish and other under-resourced languages — to international research consortia. Their engagement in explainability and human-assisted learning signals applied, real-world orientation rather than purely theoretical research.

What they specialise in

How they've shifted over time

In their first H2020 engagement (IDENTITY, 2016–2019), CENATAV worked on visual AI — computer vision applied to forensics and biometric identification from multimedia content. By 2021, their focus had shifted entirely to language and speech: the ESPERANTO project brought them into NLP, speech processing, and explainable AI for low-resource settings. This is a meaningful pivot from vision to language, though both domains share the same underlying deep learning foundations. The shift likely reflects both a strategic research repositioning and the growing global demand for robust speech and NLP tools in under-resourced language environments.

CENATAV is moving deeper into speech and language AI, with a specific angle on explainability and human-assisted learning — making them a relevant partner for any consortium working on responsible, transparent NLP systems, particularly for Spanish or other low-resource languages.

How they like to work

CENATAV participates exclusively as a third-party partner, never as a project coordinator — which is expected given Cuba's status as a non-EU, non-associated country, where MSCA-RISE is the primary participation pathway. Both their projects are MSCA-RISE exchanges, meaning their collaboration model is built around researcher mobility: sending and hosting scientists rather than leading work packages. Their 34 unique partners across 21 countries in just 2 projects indicates they engage in large, diverse international consortia typical of MSCA-RISE networks rather than tight bilateral partnerships.

Despite only two projects, CENATAV has built connections with 34 distinct partners spanning 21 countries — an unusually broad network for their project count, reflecting the multi-institutional nature of MSCA-RISE consortia. Their reach extends well beyond Europe into Latin America and other third-country regions.

What sets them apart

CENATAV is one of the very few Cuban research institutions active in EU H2020, making them a rare bridge between Latin American AI research and the European research area. Their specific expertise in low-resource language technologies is directly valuable for any NLP or speech project that needs to go beyond well-resourced languages like English or German. For consortium builders seeking MSCA-RISE third-country partners with genuine AI depth — rather than token international presence — CENATAV offers credible research capability in both computer vision and language processing.

Highlights from their portfolio

• ESPERANTO
  A 4-year MSCA-RISE project running through 2025 focused on speech research and technologies for low-resource settings, bringing together NLP, neural networks, explainability, and evaluation methodology in one large international exchange network.
• IDENTITY
  Their earliest EU engagement, on the technically demanding intersection of computer vision, multimedia forensics, and biometric person identification — demonstrating CENATAV's AI roots predate the current NLP wave.