Semantic Search
Semantic search is a search technique that understands the meaning and intent behind a query, returning results based on conceptual relevance rather than exact keyword matches.
Understanding Semantic Search
Traditional keyword search matches the exact words in your query against documents. Semantic search goes deeper by understanding what you mean. When you search for 'meeting about budget,' semantic search can find a document titled 'Q3 Financial Planning Discussion' because it understands these concepts are related. This is powered by vector embeddings that capture the meaning of text as numerical representations. Semantic search is more forgiving of different phrasing, handles synonyms naturally, and can find relevant results even when the exact terms do not appear in the document.
How GAIA Uses Semantic Search
GAIA uses semantic search powered by ChromaDB vector embeddings to find information across all your connected tools. When you ask GAIA to find something, it searches by meaning rather than keywords. You can ask 'find the email where Sarah discussed the product launch timeline' and GAIA will find it even if the email subject was 'Re: Q2 Plans.' Semantic search works across emails, tasks, documents, and notes simultaneously.
Related Concepts
Vector Embeddings
Vector embeddings are numerical representations of text, images, or other data that capture semantic meaning, enabling machines to understand similarity and relationships between pieces of information.
Knowledge Graph
A knowledge graph is a structured representation of information that organizes data as entities, their attributes, and the relationships between them, enabling machines to understand and reason about connected information.
Graph-Based Memory
Graph-based memory is an AI memory architecture that stores information as interconnected nodes and relationships, enabling rich contextual understanding and persistent knowledge across interactions.
Context Awareness
Context awareness in AI is the ability to understand the full situation surrounding a task or interaction, including who is involved, what has happened before, related projects, deadlines, and the user's preferences and patterns.
