Sony Research and AI Singapore Team Up to Develop AI Models for Southeast Asian Languages

Sony Research and AI Singapore (AISG) have signed a memorandum of understanding (MOU) to collaborate on developing large language models (LLMs) tailored to Southeast Asian languages. The partnership will center on the SEA-LION (Southeast Asian Languages In One Network) family of LLMs, which are specifically trained and tuned for the diverse linguistic landscape of the region. This initiative aims to fill the current gap in AI models representing Southeast Asian languages.

The collaboration will kick off with a focus on Tamil, a language spoken by an estimated 60-85 million people worldwide, with large populations in India and Southeast Asia. Sony Research and AISG will work together to test and refine the SEA-LION model for Tamil, while also exploring its applicability for other Southeast Asian languages. In addition to technical collaboration, the two organizations will share best practices in LLM development and research methodologies.

Hiroaki Kitano, President of Sony Research, emphasized the importance of developing AI technologies that reflect the world’s linguistic diversity. "As a global company, we recognize the significance of localization and diversity. Southeast Asia is home to over a thousand languages, and our collaboration with AISG aims to ensure that AI models cater to the needs of all populations. We’re excited to work with AI Singapore on this project, which has the potential to make AI more inclusive and representative," Kitano stated.

Sony Research’s strong presence in India will also play a key role in this partnership, particularly in leveraging their expertise in language model development for Indian languages, including Tamil.

Leslie Teo, Senior Director of AI Products at AI Singapore, expressed optimism about the collaboration. "The integration of the SEA-LION model with its Tamil language capabilities offers exciting potential for new solutions. We’re eager to contribute to its development and see how this partnership can drive innovation in multilingual AI technologies."

Both organizations see this collaboration as an opportunity to advance AI research and promote equitable technological development, ensuring that underrepresented languages are included in the rapidly growing field of AI.