Open and Trustworthy Language Models: A Consortium for European Digital Sovereignty
In a significant development for the field of artificial intelligence, the University of Turku and Silo AI, one of Europe’s largest private AI labs, have joined forces to launch a consortium focused on open and trustworthy large language models (LLMs). This initiative aims to ensure European digital sovereignty and democratize access to LLMs, with the goal of developing the world’s largest open source LLM that covers all official European languages.
Empowering European Language Models
The consortium, led by Silo AI and the TurkuNLP research group at the University of Turku, sets itself apart by combining resources to build LLMs. With access to the LUMI supercomputer, one of the largest in Europe, and a team experienced in training LLMs effectively and efficiently, the consortium is uniquely positioned to develop a family of language models, including the world’s largest open LLM.
Furthermore, the initiative is committed to ensuring that the data utilized in these models accurately represent European languages, including the English-speaking world. By collaborating with key European institutions and agencies and adhering to European regulations, the consortium aims to create language models that prioritize multilinguality and equal access to technology.
Advancing Technological Development and Digital Sovereignty
The development of open LLMs aligned with European values is crucial for maintaining digital sovereignty. By basing the underlying models on data and information that represent the citizens and organizations of the region, the consortium aims to address concerns related to regulation, data privacy, and accountability.
Peter Sarlin, CEO and co-founder of Silo AI, emphasizes the importance of trusted and secure approaches to independent base models, which can then be fine-tuned for domain-specific needs. This approach enables companies to maintain control over downstream applications and value creation, ensuring digital sovereignty while advancing technological development.
A Path to Redefine Boundaries
The TurkuNLP research group, with its extensive experience in natural language processing (NLP) and LLMs, aligns perfectly with Silo AI’s commitment to world-class research on generative AI. Together, they provide a robust foundation for pushing the boundaries of what is possible in the world of open source language models.
By offering a platform for LLM development and the ability to finetune and control models for specific domains, the consortium opens up new opportunities for companies to create value using independent, trusted, and secure base models.
Conclusion
The launch of the consortium by Silo AI and the University of Turku marks a significant step towards ensuring European digital sovereignty and democratizing access to language models. By developing the world’s largest open source LLM and prioritizing multilinguality, the consortium aims to redefine the boundaries of what is possible in the field of AI.
This initiative not only empowers European language models but also emphasizes the importance of transparency, reproducibility, and equal access to technology. By building a foundation of trusted and secure base models, the consortium paves the way for companies to leverage AI for domain-specific needs while maintaining control over downstream applications.
With the combined expertise of Silo AI, the TurkuNLP research group, and the support of key European institutions, this consortium has the potential to shape the future of language models and contribute to the advancement of AI technology.
注意
- この記事はAI(gpt-3.5-turbo)によって自動生成されたものです。
- この記事はHackerNewsに掲載された下記の記事を元に作成されています。
Consortium launched to build the largest open LLM - 自動生成された記事の内容に問題があると思われる場合にはコメント欄にてご連絡ください。