AI Agent Types

Type Function Key Tech ​​Predictor​​ Forecast outcomes Federated Learning + MCP ​​Q&A Assistant​​ Match-specific knowledge service RAG + TEE +MCP ​​Validator​​ Verify data authenticity zkProofs + SGX Enclaves

Core Technology Stack (Q&A Assistant​ AI Agent)

• Retrieval-Augmented Generation (RAG): Combines pre-trained language models with external knowledge bases to provide accurate and contextually relevant question-answering services.Every organizers collect and organize match-related Q&A information.
• Trusted Execution Environment (TEE) Encrypted Inference: Executes language model inference in an isolated and protected environment, safeguarding the privacy of knowledge bases and user queries. Especially the organizers provide Q&A information.
• Model Context Protocol (MCP) (Optional): Can be used to securely negotiate and aggregate answers or knowledge snippets among multiple Q&A Assistant​ Agents or other AI Agent function and tools, enhancing the comprehensiveness and consistency of responses.

Core Function: Q&A Assistant​ AI Agent​

Q&A Assistant​ AI Agents are designed to provide users with instant, accurate, and comprehensive knowledge services about the competition network. They can understand users' natural language queries and retrieve relevant information from vast knowledge bases to generate easily understandable answers.

• Match Knowledge Retrieval: Answers questions about specific events, such as schedules, participating teams/players, competition rules, historical data, etc.
• Real-Time Information Inquiry: Provides real-time statistics, key matchs, player performance, and other information during matches.
• Rule Explanation: Explains complex competition rules and terms, helping users better understand matchs.
• Background Knowledge Provision: Offers background information about participating teams/players, historical records, recent performance, etc.
• Multilingual Support: Capable of understanding and generating answers in multiple languages, serving a wider user base.