Thuật ngữ AI
Từ điển đầy đủ về Trí tuệ nhân tạo
Vector Indexing
Process of converting documents into vector representations to enable efficient semantic search
Semantic Search
Search methods based on understanding meaning and context rather than exact keywords
Document Chunking
Strategies for segmenting documents into fragments optimized for retrieval and generation
Query Transformation
Query reformulation and expansion techniques to improve the relevance of retrieved results
Hybrid Retrieval
Combination of vector search and keyword search to optimize precision and recall
Re-ranking Systems
Algorithms for reordering retrieved documents to maximize relevance before generation
Context Window Optimization
Optimized context window management for efficient integration of retrieved information
Multi-hop Retrieval
Iterative retrieval process requiring multiple steps to assemble complex information
Knowledge Graph RAG
Integration of structured knowledge graphs to enhance information retrieval
RAG Evaluation Metrics
Specific metrics for measuring the quality and performance of RAG systems
Adaptive Retrieval
Systems that dynamically decide when and how to retrieve information based on the query
RAG Fine-tuning
Adapting language models specifically for retrieval-augmented generation tasks
Citation Tracing
Mechanisms for tracing and attributing the sources of information generated by the RAG system
Dense Retrieval
Approaches using dense representations to capture deep semantic relationships
RAG Architecture Patterns
Architectural patterns and frameworks for implementing scalable and efficient RAG systems