Samsung Research Unveils Generative AI, Samsung Gauss
The Samsung Gauss large language model can aid in writing code and analyzing documents.
Samsung Research Unveils Proprietary Generative AI Assistant: Samsung Gauss
Samsung Research has introduced an advanced generative AI assistant named Samsung Gauss, currently serving the company's internal needs but with potential future integration into Samsung's consumer smartphones. Detailed at the Samsung AI Forum 2023 in Gyeonggi-do, Korea, Samsung Gauss aims to streamline tasks such as email composition, coding, and image generation to enhance productivity. As the field of on-device AI grows increasingly competitive, Samsung's entry will rival the likes of Google Assistant on Pixel devices and Meta’s Llama 2 on Qualcomm-powered devices starting 2024.
Understanding Samsung Gauss: An On-Device AI Marvel
Named after the renowned mathematician Carl Friedrich Gauss, Samsung’s AI assistant encapsulates a vision to leverage global knowledge and phenomena through AI to elevate consumer experiences. Samsung Gauss encompasses three distinct models, each tailored to specific tasks:
Samsung Gauss Language: A versatile language model capable of crafting emails, summarizing documents, translating content, and integrating with device control systems, potentially complementing Samsung's existing Bixby voice control.
Samsung Gauss Code: A developer's companion that supports a range of coding tasks, including code description and test case generation, powering code.i, a coding assistant platform.
Samsung Gauss Image: An artistic model designed to create and refine visual content, enhancing users' multimedia experiences.
Security at the Forefront
In light of a recent internal data leak that led to a temporary suspension of generative AI tools like ChatGPT and Google’s Bard, Samsung has underscored its commitment to security. The company’s AI Red Team is vigilantly addressing privacy and ethical considerations across the AI lifecycle, from data handling and model development to service deployment and the evaluation of AI-generated outputs.