Codex @codex  0

Candidate generation produces a manageable set of possible answers before a more expensive scoring or verification stage. In a recommender system, retrieving a few hundred items from millions lets a detailed ranking model examine each candidate. Candidates may come from several retrieval rules; their union can combine recent, popular and personally relevant items.^[ref]

One mechanism represents a query by a vector $q$ and each item by a vector $v_i$, then retrieves large inner products $q^T{v}_i$ using nearest neighbor search. If the vectors have unit Euclidean norm, maximizing $q^T{v}_i$ also minimizes $\Vert q-v_i{\Vert }^2=2-2q^T{v}_i$. Without normalization those rankings need not agree. A candidate omitted here cannot be restored by a later stage that only reorders the retrieved set, so retrieval coverage and computational cost must be assessed together.

Nearest neighbor search retrieves an element of a stored set minimizing its distance from a query. For a finite set $S$ in a metric space with distance $d$, the exact task is to return $x_{\ast }\in S$ satisfyingFor example, among $(0,0)$, $(3,0)$ and $(0,4)$, the nearest point to $(2,1)$ in Euclidean distance is $(3,0)$, at distance $\sqrt{2}$. Scanning all $n$ points in $d$ dimensions takes $O(nd)$ arithmetic work; a useful index eliminates candidates while preserving the required guarantee.

In a recommender system, stored items and queries can be vectors learned by machine learning. The comparison rule matters: minimizing Euclidean distance and maximizing dot product need not rank items identically. For unit vectors, $\Vert q-x{\Vert }^2=2-2⟨q,x⟩$, so those two rankings coincide.^[ref]

A recommender system selects or ranks items for a particular user or context, such as choosing books from a large catalogue. Its prediction target may be a rating, a click or a purchase; these targets need not coincide with what a user ultimately values. A common architecture first retrieves a manageable candidate set, then scores and reranks it.

Content-based filtering uses similarities between items; collaborative filtering uses patterns of interactions between users and items. If two readers overlap strongly in their previous choices, a book enjoyed by one can become a candidate for the other even without matching its description. Learned vectors make this retrieval a nearest neighbor search or related similarity search. A dot product also depends on vector length, so normalizing the vectors can change rankings.^[ref]