検索結果の再ランキング
候補を関連度で並べ直す
本ページの本文は英語で提供されています。タイトルと導入は日本語化されています。
この能力とは何か
Takes a query and a short list of recalled candidates, and returns a reordered ranking with relevance scores. Its division of labour differs from embedding retrieval: retrieval uses cheap vector neighbours to pull dozens from tens of thousands, whereas reranking applies a costlier model to score those dozens carefully, so it only runs over the candidate set.
技術的にどう実現するか
The typical approach is a cross-encoder: query and candidate are concatenated into one sequence and a single model outputs a relevance score; seeing both at once is more accurate than two-tower encoders, but it cannot be pre-computed and must run per pair. To control latency, a common pattern truncates with a small model first and reranks with a larger one — or simply prompts a large model to score.
代表的な製品
4関連する組織
代表的な用途
- Improving evidence quality after RAG retrieval
- Final ranking in e-commerce and content search
- Selecting the most relevant passages for a QA system
- Deduplicating and choosing among candidate answers
どう評価するか
- nDCG@k
- Discounted gain in the top-k after reranking, rewarding better placement
- MAP
- Mean of average precision over all relevant items
- MRR
- Quality of the position of the first relevant result
限界と難しさ
- Cross-encoders score each candidate separately, so latency climbs steeply with candidate count
- Judgements are unstable on out-of-distribution queries, and truncating long documents causes errors
- It can only reorder what was recalled; anything retrieval missed cannot be recovered