LLMs and Alignment. Research Talks: Dr. Shantipriya Parida

Join us for the captivating Research Talks by LLMs and Alignment, featuring Dr. Shantipriya Parida, who will delve into:

🌍 LLMs for All: Building Inclusive Models for Low-Resource Languages

In a world of approximately 7,000 languages, the advantages of Large Language Models (LLMs) are predominantly experienced by a small subset, leaving many low- and ultra-low-resource languages underserved. Addressing the critical question of how to create AI systems that truly cater to all languages, this talk will tackle the challenges and end-to-end processes involved in developing LLMs for these underrepresented languages.

Dr. Parida will present two insightful case studies:

• Odia (an Indic language) as a low-resource language
• Sámi (an Uralic language) as an ultra-low-resource language

Through these examples, we will explore techniques such as multilingual pretraining, transfer learning, tokenization strategies, and community-driven data development. The discussion will highlight key technical challenges, important design decisions, and valuable lessons learned in creating inclusive language technologies. We will conclude with a vision for scalable and equitable AI that ensures the benefits of LLMs reach all language communities.

📍 Location: Maarintie 8
🎟️ Get your tickets here!