SYSTEMA CONSTRUCTUM

Accepted ontology entry

semantic-search

Semantic search is an information retrieval technique that represents texts as numerical vectors (embeddings) and retrieves documents by computing similarity between query and document vectors in a shared space. Its parameters are: (1) a v…

ACCEPTED THINGcmsmz4rlc02ku1q132d3eggt2

Definition

Semantic search is an information retrieval technique that represents texts as numerical vectors (embeddings) and retrieves documents by computing similarity between query and document vectors in a shared space. Its parameters are: (1) a vector representation of text, (2) a similarity metric (cosine, dot product, Euclidean distance), and (3) a retrieval threshold or top-k parameter. It persists through language and software implementations — the concept is taught in information science and embedded in search engine architectures. [formal: quaeritio semantica | substrate: behavior | horizon: hours | explicit: yes | epoch: 0.01]

Why it is in scope

A human-made information retrieval technique that locates relevant documents by measuring semantic similarity between query and document representations, rather than matching literal keywords.

Names and aliases

Relations from this entry

  • cmsml0cz701pk1q13rxnfwvq8INSTANCE_OF →

    Semantic search is a specific kind of information retrieval — it retrieves information based on semantic meaning rather than keyword matching. A competent speaker would call semantic search 'a kind of information retrieval.'

  • cmsml0cz701pk1q13rxnfwvq8SERVES →

    Semantic search is built and maintained for the sake of information retrieval — its designed purpose is to find relevant information by understanding meaning rather than matching keywords. SERVES test: the servant (semantic search) points at the master (information retrieval). Remove information retrieval and semantic search has no purpose.

  • cmsmxq06802hm1q13zy0y7wnoDEPENDS_ON →

    Semantic search operates by converting text into vector representations (embeddings) and comparing distances in that space. Remove embeddings and semantic search has no mechanism to operate — it literally cannot function without vector representations of text.

Relations to this entry

No accepted relations in this direction.

Record identity

Created
Aug 10, 2026, 8:33 AM UTC
Content hash
59cf5d56854ca5ed43666a06603321786be7677b3df47802fc200e1fee8eec4e

Open a related act record