EdTech Tracker

← Coding lens

reciprocal rank fusion

technique judged 100% likely a named technique or architecture

First seen
2024-06-26
Independent sources
8
Articles
17
Last 4 weeks vs prior
0 vs 3

See how it has spread across media tiers

Mentions over time

Articles per week (by publish date) using the term, in any spelling.
01.532024-W262026-W152026-W40
Table view
WeekArticlesSources
2024-W2611
2025-W2311
2026-W0411
2026-W0911
2026-W1121
2026-W1311
2026-W1511
2026-W2411
2026-W3233
2026-W3511
2026-W3711
2026-W4032

In their words

Fuse the results with reciprocal rank fusion or a weighted score, then rerank.

Reciprocal rank fusion combines ranking positions instead of assuming that scores from different retrieval methods are directly comparable.

The fix is dense neural embeddings combined with traditional keyword search (BM25) in a hybrid approach, fused using Reciprocal Rank Fusion(RRF) rather than an arbitrary weighted average.

Multi-Query fans out 3–5 paraphrases in parallel and fuses results with Reciprocal Rank Fusion (RRF).

For further optimization, you might consider using a hybrid approach, where you identify pages with charts to implement as multi-vectors and the remaining corpus as text, and then use an RRF-style approach to merge results at query time.

| Layer | Baseline | Qdrant + Minima | |---|---|---| | Agent loop | Plan, retrieve, check evidence, refine once if needed, answer with citations | Identical prompt, tool schema, stopping rule, and call limit | | Retrieval | Qdrant dense retrieval, indexed payload filters, top 16 | Dense plus BM25 sparse, RRF fusion, ColBERT-style late-interaction reranking, the same indexed payload filters, top 8

Methodology ​ As in our first Search Mode Benchmarking blog, we compare against Weaviate's Hybrid Search, combining BM25 with vector search over Snowflake Arctic 2.0 embeddings, fused with Reciprocal Rank Fusion (RRF).

The next time that same tool fails with an invalid-query error, the recall hook searches the tool’s own pattern library, matching hybrid on both the failing query text and the error message, against the vector index and the fulltext index, fused with reciprocal rank fusion, always scoped to the failing tool.