Samsung Electronics recently hosted the Samsung Developer Conference Korea 2024 (SDC24 Korea), a virtual showcase of its latest breakthroughs in software, artificial intelligence, IoT, and open-source innovation. Among the highlights was the unveiling of Gauss2, Samsung’s next-generation AI model, designed to redefine efficiency and versatility across diverse applications.
Building upon the foundation of its predecessor, Samsung’s first in-house AI model, Gauss2 boasts enhanced capabilities for managing multimodal data, including text, code, and images. To cater to varied needs, Gauss2 is offered in three versions:
- Compact Model: Tailored for devices with limited computing power, enabling seamless on-device AI experiences.
- Balanced Model: A middle ground that optimises speed, efficiency, and performance, ideal for general-purpose tasks.
- Supreme Model: Incorporating advanced “mixture of experts” (MoE) technology, it achieves high performance while reducing computational overhead.
“Gauss2 is already transforming our internal productivity, and we’re excited to bring its benefits to our products for enhanced convenience and personalisation,” said Paul Kyungwhoon Cheun, President and CTO of Samsung’s Device eXperience (DX) Division.
Samsung Unveils Next-Gen Gauss2 AI Model
Gauss2 supports up to 14 spoken and programming languages, equipped with custom stabilisation techniques and an advanced tokeniser for improved efficiency in tasks like multilingual translations and complex coding. Internally, the AI model powers code.i, Samsung’s AI-driven coding assistant. Since its release in December, usage has surged fourfold, with 60% of software developers in Samsung’s Korean DX Division adopting the tool.
Additionally, the Samsung Gauss Portal, a conversational AI assistant, helps employees automate tasks such as document summarisation, translations, and email drafting. The portal has also been extended to international subsidiaries, while Gauss-enabled tools are streamlining customer service by categorising and summarising interactions.
Looking ahead, Samsung plans to expand Gauss2’s capabilities in natural language processing, data interpretation, and image generation. Under its “AI for All” initiative, the company envisions Gauss2 enriching its entire product ecosystem, offering smarter, more personalised user experiences. By integrating AI with knowledge graph technology, Samsung aims to deliver unprecedented levels of customisation, revolutionising the way users interact with technology.
