DeepTingji has officially launched SWAN, an embodied intelligence architecture specifically designed for outdoor companion robots. This architecture features a real-time decision-making chain, a hybrid memory database, and a secure data flywheel system, creating a closed loop for decision-making, memory, and iteration. SWAN will first be implemented in DeepTingji's consumer-grade dual-wheeled outdoor companion robot, Rovar, with an expected market launch by the end of the year. Additionally, the National Supercomputing Internet has integrated the Kimi K3 model API, allowing developers to access its 2.8 trillion parameter capabilities with simplified setup and compatibility with OpenAI and Anthropic interfaces. AI
IMPACT Enables more sophisticated outdoor robotics and simplifies access to large language models for developers.
RANK_REASON The cluster describes the launch of a specific AI architecture for robots and the integration of an existing large language model API into a supercomputing service, which falls under AI tooling and infrastructure.
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
