{"id":10899,"date":"2025-10-27T18:31:27","date_gmt":"2025-10-27T17:31:27","guid":{"rendered":"https:\/\/eurocc.nscc.sk\/?p=10899"},"modified":"2025-10-27T18:31:27","modified_gmt":"2025-10-27T17:31:27","slug":"webinar-amalia-smerom-k-multimodalnemu-llm-pre-europsku-portugalcinu","status":"publish","type":"post","link":"https:\/\/eurocc.nscc.sk\/en\/webinar-amalia-smerom-k-multimodalnemu-llm-pre-europsku-portugalcinu\/","title":{"rendered":"Webinar am\u00e1lia: Towards a Multimodal LLM for European Portuguese"},"content":{"rendered":"
\n
\n
\n
Webinar am\u00e1lia: Towards a Multimodal LLM for European Portuguese<\/h5>\n\n\n\n

Join us for an inspiring session on the development of am\u00e1lia, Portugal\u2019s large language model designed to bring the richness of European Portuguese into the new era of multimodal artificial intelligence.<\/strong><\/p>\n\n\n\n

This webinar will feature Prof. Jo\u00e3o Magalh\u00e3es from NOVA LINCS, Universidade NOVA de Lisboa, who will present the goals, architecture, and progress of this national AI initiative. The talk will explore how am\u00e1lia combines text, speech, image, and video understanding, and how it contributes to building culturally aligned and trustworthy AI systems for public, academic, and enterprise use.<\/p>\n\n\n\n

Date and Time:<\/strong>
Wednesday, November 12th, 2025 | 10:00 AM CEST (9:00 PT)
Online | Free Registration<\/p>\n\n\n\n

\n
REGISTRATION<\/a><\/div>\n<\/div>\n<\/div>\n\n\n\n
\n
\n
\"\"<\/a><\/figure>\n<\/figure>\n<\/div>\n<\/div>\n\n\n\n

This webinar is organized by the Slovak National Supercomputing Centre<\/strong> as part of the EuroCC project (National Competence Centre \u2013 NCC Slovakia<\/a><\/strong>) in cooperation with NCC Portugal<\/strong> , within the LLM Webinar Series connecting high-performance computing with artificial intelligence, culture, and innovation. The webinar will be held in English.<\/p>\n\n\n\n

The webinar will be held in English.<\/p>\n\n\n\n

Abstract:<\/strong><\/p>\n\n\n\n

am\u00e1lia<\/a><\/strong> is a government-backed large language model focused on European Portuguese, designed to capture linguistic nuances, cultural context, and multimodal capability across text, speech, image, and video. The project aims for high-impact applications in public administration, research, and industry, with a strong emphasis on trust, alignment, and data sovereignty.<\/p>\n\n\n\n

This presentation outlines the process and key methodologies employed by the am\u00e1lia LLM Team at NOVA University and Instituto Superior T\u00e9cnico \/ UL, with a focus on utilizing European HPC resources such as the Marenostrum 5 (MN5) and Deucalion supercomputers. Prior Portuguese-language initiatives (like Gl\u00f3rIA) and European initiatives (such as EuroLLM) inform benchmarks and evaluation and help define the roadmap from beta to public release.<\/p>\n\n\n\n

The core development pipeline is explored through three key dimensions:<\/strong>
1\ufe0f Initial data preparation and training, involving an extensive, multi-month process of transforming noisy HTML and PDFs into high-quality raw texts, followed by tokenization and core training methodologies in language modeling and instruction tuning.
2\ufe0f. Model alignment, achieved through reinforcement learning approaches, Direct Preference Optimization (DPO) and Group Relative Policy Optimization (GRPO) with Verifiable Rewards (VR) to ensure safer and more trustworthy responses.
3\ufe0f. Infrastructure setup for advanced RLVR (GRPO) training on MN5, which uses inference nodes for sampling and training nodes for collecting samples and running Verifiable Rewards, highlighting the complexity of configuring multiple custom environments (e.g., mathematics, programming, biology).<\/p>\n\n\n\n

The talk concludes with key insights into the computational and methodological rigor required to efficiently develop state-of-the-art LLMs, positioning this work at the forefront of Europe\u2019s innovation path in AI.<\/p>\n\n\n\n

Speaker:<\/strong><\/p>\n\n\n\n

Prof. Jo\u00e3o Magalh\u00e3es<\/strong> \u2013 CMU Portugal co-Director; Head of the Multimodal Systems Group, NOVA LINCS, Universidade NOVA de Lisboa<\/p>\n\n\n\n

\"\"

<\/figcaption><\/figure>\n\n\n\n

Prof. Jo\u00e3o Miguel da Costa Magalh\u00e3es is a Full Professor in the Department of Informatics at Universidade NOVA de Lisboa and a senior researcher at NOVA LINCS. He serves as co-Director of the CMU Portugal Program and leads the Multimodal Systems Group. His research focuses on vision-language models, multimodal learning, and AI systems for semantic multimedia. He earned his PhD from Imperial College London and has been a key figure in Portugal\u2019s AI and digital innovation ecosystem.<\/p>\n\n\n\n

<\/p>\n\n\n\n

Topics Include:<\/strong><\/p>\n\n\n\n